Editor's Note: As the "nerve endings of the information age," sensors have permeated every critical field of the socio-economic landscape. Since October 2025, China Electronics News has invited Guo Yuansheng, Deputy Director of the Science and Technology Committee of the Jiusan Society and Executive Vice Chairman of the China Sensor and IoT (Internet of Things) Industry Alliance, to launch the column "Guo Yuansheng Explains Sensors." Focusing on eight major fields and scenarios—electric power, major equipment, intelligent manufacturing, smart agriculture, smart healthcare and big health, smart home appliances and consumer electronics, urban security, and low-altitude economy—the column has published articles such as "Sensors on the Power Generation Side: The Cornerstone of Stable Operation in New Power Systems" and "Three Core Development Directions Clarified for Energy Storage Sensors," which have received widespread attention and high praise from readers. This issue features the intelligent manufacturing section, focusing on sensors for intelligent manufacturing, elaborating on their application prospects, industrial status, and future recommendations to build industrial consensus and promote industrial development.
Given the global context of industrial digitalization and the digitalization of industries, intelligent manufacturing has emerged as a critical force driving industrial upgrading and enhancing national core competitiveness. Through the deep integration of AI (Artificial Intelligence), information technology, and automation technology with manufacturing processes, it realizes the digitalization, intelligence, flexibility, and efficiency of production. As the only functional devices for data acquisition and the "electronic senses and skin," sensors play a vital role in this process. They can dynamically and in real-time perceive various physical, chemical, and biological quantities during production, converting this information into processable data signals, thereby providing a solid foundation for the decision-making and control of intelligent manufacturing systems.
Whether it is material distribution and status detection at the front end of industrialization, real-time monitoring of production efficiency, product quality, and equipment operation during the intermediate process, or the management of cargo storage environments and transportation status in backend warehousing and logistics, all rely on the support of sensors. As intelligent manufacturing continues to iterate, higher requirements will be placed on the performance, functionality, and integration level of sensors. Therefore, how to further enhance the industrialization capability and application level of industrial discrete sensors has become an urgent issue to be addressed.
Core Technology Evolution and Category Analysis of Intelligent Manufacturing Sensors
Currently, industrial discrete sensors have formed a complete product system comprising eight major sensing technologies: acoustic, force, optical, gas, magnetic, temperature and humidity, RFID, and biological. They are engaged in technological innovation and comprehensive competition around four main lines: precision and reliability upgrading, multi-parameter integration, intelligent modules, and interconnected nodes.
Mechanisms, Characteristics, Indicator Systems, and Product Forms of the Eight Major Sensing Technologies
The sensing mechanism is the technical core and functional carrier of sensors. The performance indicators and product forms formed by different sensing mechanisms and technologies determine their adaptability to application scenarios. The following systematically analyzes the technical characteristics of the eight major sensing mechanism types from three dimensions: core performance indicators, product structure and form, and typical application scenarios:
| Sensing Type | Core Performance Indicators | Product Structure and Form | Typical Application Scenarios | Technical Principles and Development Status |
| Acoustic Sensing Devices and Technologies | Sensitivity ≤20Pa, response time <1ms, frequency response range 20Hz-1MHz, detection distance 0.1-50m | Piezoelectric (Lead Zirconate Titanate PZT material) / Capacitive, ultrasonic probes (diameter 10-50mm), array acoustic sensors (8-32 channels) | AGV obstacle avoidance, equipment abnormal noise monitoring, weld seam detection, fluid flow measurement | Acoustic signal conversion is realized based on the piezoelectric effect or capacitance change. Array types improve positioning accuracy through multi-channel fusion. Mainstream products have achieved digital output, and some high-end products integrate noise suppression algorithms. |
| Force Sensing Devices and Technologies | Linearity error <3%, repeatability ±0.1%FS, overload stability 150% FS, resolution 0.001% FS | Strain gauge type (metal strain gauges) / Piezoelectric, tactile array sensors (minimum unit 0.5×0.5mm, sensing density 400 points/cm²), torque sensors | Robot gripping force control, stamping process pressure monitoring, assembly force feedback, equipment load detection | Strain gauge type: Metal deformation is converted into resistance change. Piezoelectric type: Based on charge output of piezoelectric crystals. Tactile array: Miniaturization is achieved using MEMS (Micro-Electro-Mechanical Systems) technology, widely used in collaborative robots. |
| Optical Sensing Devices and Technologies | Spatial resolution 1.7cm, field of view 70°×70°, detection rate >99%, response time <50μs | LiDAR (128×128 pixel array, ranging distance 0.5-200m), vision sensors (CCD/CMOS chips, 1 million to 20 million pixels), optical fiber sensors (transmission distance 0-10km) | 3D imaging, part positioning (LiDAR), PCB (Printed Circuit Board) quality inspection, surface defect detection (vision), long-distance signal transmission (optical fiber sensors) | LiDAR is based on the Time of Flight (ToF) or triangulation principle. Vision sensors achieve target recognition through image pixel analysis. Optical fiber sensors perceive physical quantities using light refraction/reflection changes, with outstanding anti-electromagnetic interference capabilities. |
| Gas Sensing Devices and Technologies | Detection limit at the ppm level (some at the ppb level), response time <10s, attenuation after 10 million cycles <5%, operating temperature -40℃~85℃ | Electrochemical (three-electrode system) / Infrared gas, array gas detectors (simultaneous detection of 4-8 gases), flexible gas sensors | Workshop harmful gas monitoring (CO, NO₂, VOCs), lithium battery production environment control (O₂, H₂), chemical raw material leak detection | Electrochemical sensors generate current signals through gas reactions. Infrared sensors are based on the characteristic absorption spectra of gas molecules. Array products can achieve cross-detection of multiple gases, reducing the false alarm rate. |
| Magnetic Sensing Devices and Technologies | Resolution at the nT level, linear range ±10mT, temperature stability -40℃~125℃, response time <1μs | Hall sensors (SiGe material), magnetoresistive sensors (GMR/TMR thin films), fluxgate sensors, magnetic encoders | Motor speed detection, metal part flaw detection, position positioning, current measurement | Utilizing the Hall effect, magnetoresistive sensors are based on magnetoresistance changes. Fluxgate sensors convert weak magnetic fields into voltage signals through the nonlinear magnetization characteristics of the iron core, enabling nT-level weak magnetic field detection. |
| Temperature and Humidity Sensing Technologies | Temperature accuracy ±0.1℃, humidity accuracy ±2% RH, long-term stability >5 years, response time <8s | Thermistors (NTC/PTC ceramic materials), capacitive humidity sensors (polymer films), integrated modules (I²C/SPI interfaces) | Warehousing environment monitoring, equipment temperature monitoring, process temperature control, cold chain logistics | The resistance value of thermistors changes with temperature. Capacitive humidity sensors change capacitance through polymer moisture absorption. Integrated modules have built-in calibration algorithms and directly output digital signals, facilitating system integration. |
| RFID Devices and Technologies | Read/write distance 0-10m (passive) / 0-100m (active), recognition speed >50 tags/second, strong anti-interference (ISO 18000-6C) | Passive/active tags (size 3mm×3mm~100mm×100mm), UHF readers (operating frequency 860-960MHz), integrated modules (embedded with MCU (Microcontroller Unit) chips) | Material traceability, inventory counting, asset tracking, personnel positioning | Data read/write is realized based on electromagnetic coupling or electromagnetic wave propagation. Passive tags rely on the energy supply from readers. Active tags have built-in batteries and can actively send signals. UHF RFID supports simultaneous recognition of multiple tags. |
| Biological Sensing Technologies | Specific cross-reaction rate <5%, detection time <30min, sensitivity at the ng level, sample requirement <10μL | Biochips (DNA/protein chips, 100-1000 detection sites), immunosensors (specific antigen-antibody binding), enzyme electrode sensors | Biopharmaceutical raw material detection, food quality traceability, microbial contamination monitoring | Utilizing the specific recognition ability of biological molecules, biological reactions are converted into electrical or optical signals. Biochips achieve parallel detection of multiple indicators, and enzyme electrode sensors are widely used in food freshness detection. |
The eight major sensing technologies each have their unique characteristics: Acoustic sensing technology responds rapidly in dynamic detection, with a huge industrial scale and clear product supporting markets; force sensing device process technologies are scattered, but they have wide applications and huge market demand, showing great future potential; optical sensing devices have high technical difficulty and core technologies are hard to master, but they will dominate the future direction of visual perception and 3D imaging technology, with obvious industrial characteristics; gas sensing technology and industry are highly dispersed, and application innovations in security and medical fields are imminent; magnetic sensing technology and industry have high concentration, with strong market supporting demand; temperature and humidity sensing has broad prospects in basic environmental monitoring, smart agriculture, and the big health industry; RFID technology applications are growing rapidly across industries; and biological sensing has clear applications in biomedicine, the food industry, and health monitoring.
Four Main Lines of Technological Innovation and Development
As the requirements for sensing technology in intelligent manufacturing steadily improve, four major innovation lines have been formed: precision and reliability upgrading, multi-parameter integrated integration, deep intelligent empowerment, and interconnected nodes, driving discrete sensors to leap from "single perception" to "intelligent perception" and "microsystems."
1. Precision and Reliability Upgrading: Breaking Through the Adaptability to Extreme Working Conditions
(1) Leapfrog development from analog to digital: Traditional analog sensors suffer from issues such as signals being easily interfered with and limited precision. Digital sensors achieve direct signal digitization through built-in AD converters (12-bit to 24-bit). For displacement detection sensors, the detection accuracy has improved from ±0.1mm in analog types to ±0.01mm in digital types;
(2) Breakthroughs in environmental adaptability: High temperature, high humidity, and strong electromagnetic interference in industrial scenarios are the main challenges for sensor applications. Currently, high temperatures can be addressed using ceramic packaging and high-temperature-resistant materials (such as AlN ceramics and SiC semiconductors), enabling stable operation in environments from -40℃ to 150℃; high humidity environments are managed through waterproof and breathable membrane technology, keeping the error rate within 3% under 95% RH (non-condensing) conditions; anti-electromagnetic interference adopts shielded packaging and differential signal transmission design, with a signal distortion rate of <1% under 1000V/m electromagnetic radiation, meeting the needs of complex scenarios such as automotive manufacturing and power equipment.
2. Multi-parameter Integrated Integration: Reducing Deployment Costs and Space Occupation
(1) Single-component multi-physical quantity detection: Traditional single parameters require the combined deployment of multiple devices. Composite sensors achieve "one device, multiple measurements" by integrating various sensing units. For example, the "temperature-pressure" composite integrates thermistors and piezoresistive sensing units, reducing the volume by 50% and deployment costs by over 30%, and has been applied to engine block detection; the "humidity-gas" integrated detector simultaneously monitors environmental humidity and harmful gas concentrations, avoiding the impact of humidity changes on gas detection accuracy, and achieves synchronous monitoring of H₂ and humidity in lithium battery workshops;
(2) Array-based perception: Through high-density sensing unit arrays, multi-dimensional and large-area perception is realized. For example, a pressure distribution detection system with 262,000 sensing points (1024×256 array) can collect spatial distribution data of object contact pressure in real time, with a resolution of 0.1mm, applied in robot flexible gripping and automotive seat comfort testing; a high-density tactile sensing matrix (minimum unit 0.1×0.1mm) is manufactured using MEMS technology, achieving human skin-level tactile perception and supporting fine operations of humanoid robots.
3. Deep Intelligent Empowerment: From Data Acquisition to Intelligent Cognition
(1) Edge computing integration: Sensors have built-in microprocessors (MCU (Microcontroller Unit)) to realize local data preprocessing, filtering, and feature extraction, uploading only valid data to the cloud, reducing cloud transmission pressure by 50%. For example, industrial vibration sensors have built-in FFT (Fast Fourier Transform) algorithms to locally analyze vibration frequency characteristics and upload only abnormal frequency data, reducing data transmission volume by 90%;
(2) AI (Artificial Intelligence) algorithm embedding: Deep learning and machine learning algorithms are integrated into modules to realize intelligent data analysis and decision-making. For example, adopting an "industrial-grade optoelectronics + deep learning" solution, image data is collected through vision sensors, and defect recognition algorithms are run locally, increasing the semiconductor defect detection rate by 8% and reducing the false alarm rate to below 0.5%; the built-in self-diagnosis function analyzes its own working status through AI algorithms, with equipment failure warning accuracy exceeding 85%, predicting sensor failure risks 30 days in advance;
(3) Microprocessor integration: Deep integration with MCU and DSP (Digital Signal Processor) realizes real-time data analysis and decision feedback, with response delays as low as the millisecond level. For example, integrating MCU and PID control algorithms to adjust gripping force in real time during robot assembly, with a response delay of <1ms, ensuring no damage during precision part assembly.
4. Interconnection Breakthroughs: Supporting Industrial IoT (Internet of Things) Collaboration
(1) Wireless communication adaptation: Addressing the large-scale deployment needs of the Industrial IoT, sensors support wireless protocols such as LPWAN (LoRa/NB-IoT), Bluetooth 5.0, and Wi-Fi 6, realizing low-power long-distance transmission. The LoRa protocol has a communication distance of 3-10km, with an operating current as low as 10μA (sleep current 1μA), suitable for large-scale deployment scenarios such as factory workshops and warehousing logistics; NB-IoT sensors rely on operator networks, requiring no self-built base stations, with stable data uploading, and have been applied in cross-regional asset tracking;
(2) Industrial Ethernet compatibility: To meet real-time control needs, sensors adapt to industrial Ethernet protocols such as PROFINET, EtherCAT, and Modbus TCP, with data transmission delays ≤1ms. EtherCAT protocol sensors support distributed clock synchronization with a synchronization accuracy of 100ns, supporting multi-device collaborative control, realizing synchronous actions and data interaction of 128 robots in automotive welding production lines; PROFINET sensors seamlessly dock with PLC and MES systems, increasing the data interaction rate to over 95% and breaking "data silos."
In-depth Analysis of Digital Scenario Applications in the Full Process of Intelligent Manufacturing
The value of sensors is ultimately realized through scenario applications. In the full process of intelligent manufacturing, the three major links—front-end material distribution and warehousing logistics, mid-end manufacturing processes and quality control, and back-end equipment operation and maintenance and energy efficiency optimization—form a closed-loop system of "perception-decision-execution." The collaborative application of the eight major sensing technologies promotes the comprehensive improvement of production efficiency, product quality, and operation and maintenance levels.
Front End: Material Distribution and Warehousing Logistics
Material distribution and warehousing logistics are the "supply chain cornerstone" of intelligent manufacturing, and their efficiency and accuracy directly affect the production rhythm. Through functions such as environmental perception, position positioning, and status monitoring, sensors realize "precise distribution, safe storage, and dynamic traceability" of materials, supporting flexible production and lean management.
1. Core Sensor Combinations
(1) Optical sensing (LiDAR) + acoustic sensing (ultrasonic): LiDAR realizes 3D environment modeling and obstacle detection. Ultrasonic sensors make up for the short-range blind spots of LiDAR. The two work together to realize precise obstacle avoidance and path planning for AGVs (Automated Guided Vehicles);
(2) Temperature and humidity + gas: Real-time monitoring of warehousing environment temperature and humidity (accuracy ±0.1℃/±2% RH), combined with gas sensors to detect harmful gases (such as CO, VOCs) and flammable and explosive gases (such as H₂, methane), preventing risks of material moisture, oxidation, and fire;
(3) RFID + force sensing: RFID tags record full life cycle information such as material model, batch, and shelf life. Combined with the pressure during material loading and unloading (range 0-500kg), it avoids damage to fragile and precision materials.
2. Typical Application Cases
(1) AGV navigation: The AGV navigation system at Tesla's Shanghai Gigafactory adopts a collaborative perception solution of ultrasonic sensors and pixel array LiDAR. Ultrasonic sensors have a detection distance of 0.1-5m and a response time of <100μs, responsible for short-range obstacle avoidance. LiDAR has a ranging distance of 0.5-100m and a spatial resolution of 1.7cm to realize global path planning. The navigation error is ≤±5mm, and the obstacle avoidance response time is <100ms. The AGV travel speed reaches 2m/s, increasing material distribution efficiency by 40% and reducing costs by 60% compared to traditional manual distribution;
(2) Precise distribution: A semiconductor factory uses flow sensors (accuracy ±0.5%FS) and pressure sensors (range 0-1MPa, linearity error <0.1% FS) to link and control material delivery pipelines. Flow sensors monitor the material delivery rate in real time, and pressure sensors feedback pipeline pressure. The pump speed is dynamically adjusted through the PLC system. The material distribution accuracy reaches 99.9%, avoiding production interruptions caused by insufficient or excessive material supply. The distribution efficiency increases by 30%, saving 12 million RMB in production costs annually; (3) Inventory counting: An automotive parts enterprise adopts UHF RFID batch scanning technology. Passive RFID tags are pasted on warehouse shelves and material boxes, with a read/write distance of 0-5m and a recognition speed of >50 tags/second. AGVs equipped with RFID readers automatically cruise in the warehouse, covering 1,000 shelves in a single inventory count. The inventory efficiency increases by 5 times, and the error rate drops to below 0.1%, realizing dynamic visual management of inventory. The system can provide real-time warnings for material shortages and expiration risks, increasing the material turnover rate by 25% and reducing inventory backlog costs by 30%;
(4) Cold chain logistics monitoring: A pharmaceutical enterprise deployed temperature and humidity sensors, a GPS positioning module, and an NB-IoT communication module in its vaccine cold chain logistics vehicles. The operating temperature is -20℃~25℃, with a sampling frequency of 1 time/minute. Sensors collect temperature and humidity data in the carriage in real time and upload it to the cloud platform via the NB-IoT network. If the temperature deviates from the safe range of 2-8℃, the system immediately triggers audible and visual alarms and SMS notifications to ensure vaccine quality. This solution reduces the cold chain logistics temperature exceedance rate from 3.2% to 0.2%, and reduces vaccine loss costs by 80%.
Mid End: Manufacturing Processes and Quality Control
Process digitalization and quality control are the core of intelligent manufacturing. Through functions such as process parameter monitoring, product defect detection, and safety protection, it realizes "real-time control, precise detection, and safe and efficient" production processes, promoting the transformation of manufacturing from "scale production" to "personalized customization and precise configuration."
1. Core Sensor Combinations
(1) Optical sensing (3D laser) + vision: 3D laser sensors realize 3D size measurement and positioning of parts (accuracy ±0.01mm). Vision sensors perform surface defect detection and character recognition. The two work together to support precision assembly and quality inspection;
(2) Force sensing + magnetic sensing: Real-time monitoring of pressure/torque parameters in processes such as stamping, welding, and assembly (resolution 0.01N). Magnetic sensing technology realizes motor speed detection and metal part flaw detection, ensuring process stability and product consistency;
(3) Safety light curtain (optical sensing) + vibration sensing: Safety light curtains form an infrared protective barrier (response time <10ms) to prevent personnel from mistakenly entering high-risk operation areas; vibration sensors monitor equipment operation vibration (frequency range 10Hz-10kHz) to realize equipment status warning and fault diagnosis.
2. Typical Application Cases
(1) Automotive welding quality control: The body welding production line of a NEV (New Energy Vehicle) factory uses 3D laser sensors (128×128 pixel array, ranging accuracy ±0.1mm) to detect weld seam gaps. The sensors are installed at the end of welding to collect weld seam cross-section data in real time and compare them with standard parameters. If the deviation exceeds 0.05mm, the system immediately adjusts the welding current and speed, with a defect detection rate of over 99%. Meanwhile, magnetoresistive sensors monitor the motor speed of welding robots (resolution 1rpm) to ensure stable welding speed, increasing the qualified rate of whole vehicle welding quality from 98.5% to 99.8%;
(2) Electronic component assembly: The chip mounting production line of a smartphone factory adopts collaborative control of vision sensors (20 million pixels, positioning accuracy ±0.01mm) and force sensors (range 0-50N, resolution 0.01N). Vision sensors identify the chip pad positions on the PCB (Printed Circuit Board) and guide the mounting head for precise positioning. Force sensors monitor the mounting pressure in real time to avoid chip damage caused by excessive pressure. This solution increases the first-pass yield of electronic component assembly to 99.5%, reduces rework costs by 70%, and increases production efficiency by 25%;
(3) Stamping workshop safety and process optimization: A home appliance enterprise's stamping workshop deployed safety light curtains (protection height 1.2-2.5m, response time <5ms) and vibration sensors (range 0-50g, frequency response 10Hz-10kHz). When personnel approach the stamping mold, the safety light curtain immediately triggers equipment shutdown, reducing the accident rate by 80%; collecting press operation vibration data and analyzing equipment wear status through AI algorithms, the predictive maintenance accuracy reaches 85%, equipment downtime is reduced by 40%, and the scrap rate of stamped parts drops from 1.8% to 0.5%;
(4) Defect detection: An electronic manufacturing enterprise uses vision and optical fiber sensors to collaboratively detect PCB defects. Vision sensors (10 million pixels) capture surface images of the PCB to identify defects such as circuit short circuits, open circuits, and pad offsets; optical fiber sensors (detection accuracy 0.01mm) detect PCB thickness and hole diameters to ensure compliance with design requirements. This solution achieves a detection speed of 300mm/s, a defect detection rate of over 99.2%, which is 10 times more efficient than manual detection, and reduces the false detection rate to below 0.3%.
Back End: Equipment Operation and Maintenance and Energy Efficiency Optimization
Equipment operation and maintenance and energy efficiency optimization are the guarantees of intelligent manufacturing. Through functions such as equipment status monitoring, energy consumption data collection, and fault diagnosis, sensors realize "predictive maintenance, energy efficiency improvement, and green production," reducing operation and maintenance costs and energy consumption.
1. Core Sensor Combinations
(1) Vibration + acoustic sensing + temperature sensing: Collecting vibration signals of equipment rotating parts (motors, bearings, gears) (frequency 10-1000Hz), silicon microphones monitor noise changes and abnormal sounds (frequency 20Hz-20kHz), combined with monitoring part temperature (accuracy ±0.1℃), the two work together to realize equipment health status evaluation;
(2) Current + temperature and humidity sensing: Current sensors (accuracy ±0.5%) monitor the power consumption of production equipment. Temperature and humidity sensors collect workshop environmental parameters. Energy efficiency optimization is realized by dynamically adjusting equipment operating loads and air conditioning systems;
(3) Infrared + acoustic sensing: Infrared temperature measurement range is -40℃~500℃, with an accuracy of ±1℃, detecting abnormal surface temperatures of equipment. Acoustic sensing captures abnormal sounds during equipment operation, working together to realize precise fault diagnosis.
2. Typical Application Cases
(1) Predictive maintenance of rotating equipment: A steel enterprise's blast furnace fan system deployed vibration sensors (range 0-20g, resolution 0.001g) and temperature sensors (operating temperature -40℃~200℃). Data is collected every 10 seconds, and the vibration frequency spectrum and temperature change trends are analyzed through edge computing modules, combined with AI algorithms to predict equipment failures. This makes the fault prediction accuracy exceed 85%, reduces maintenance costs by 40%, and avoids production interruption losses caused by sudden failures;
(2) Energy efficiency optimization: A large energy storage base uses energy consumption sensors (monitoring accuracy ±0.2%) and temperature and humidity sensors (accuracy ±0.1℃/±2%RH) to optimize operation efficiency. Energy consumption sensors collect the charging and discharging power and grid loss data of energy storage battery packs in real time, and temperature and humidity sensors monitor environmental parameters in the battery cabin. The system dynamically adjusts the charging and discharging current and optimizes the air conditioning cooling power, achieving an overall power saving rate of 15% and saving over 8 million RMB in electricity costs annually;
(3) Equipment fault diagnosis: An automotive parts factory uses infrared sensors (temperature measurement range -20℃~300℃) and acoustic sensors (sensitivity ≤20Pa) to diagnose motor faults. Infrared sensors monitor the motor bearing temperature and trigger a warning if the temperature exceeds 80℃; acoustic sensors capture abnormal sounds during motor operation and identify fault types such as bearing wear and rotor imbalance through spectrum analysis. This solution reduces fault localization time by 60%, increases maintenance efficiency by 50%, and extends the motor's Mean Time Between Failures (MTBF) from 8,000 hours to 12,000 hours;
(4) Workshop energy efficiency monitoring: An electronic factory deployed 1,200 energy consumption sensors (covering production equipment, lighting, air conditioning, etc.) to collect power consumption data of each link in real time and analyze energy consumption distribution through the Industrial Internet platform. The system identifies problems such as excessive cooling of the air conditioning system and high standby power consumption of equipment. By optimizing temperature settings (adjusting from 22℃ to 24℃) and turning off idle equipment, the overall energy consumption of the workshop is reduced by 12%, saving 6 million RMB in energy costs annually.
Industrialization Problems, Difficulties, and Breakthrough Solutions
Although China's sensor technology has made progress in intelligent manufacturing applications, it still faces many challenges in high-end technology breakthroughs, system integration collaboration, and cost control. To address these issues, it is necessary to promote high-quality industrial development through three major paths: domestic substitution, standardization construction, and ecological cooperation.
Three Core Challenges
1. Technological Bottlenecks: "Choke Points" in High-end Fields
(1) Dependence on imports for sensing chips and core components: The core of sensors is MEMS chips, high-end ASIC circuits, and sensing materials, with a low domestic production rate. Data shows that the domestic production rate of MEMS chips is about 35%, and that of high-end ASIC circuits is less than 20%; the domestic production rate of BAW filters is about 8%, which restricts the development of the industrial high-end sensor industry;
(2) Obvious performance gap: There is a significant gap with international advanced levels. For example, the drift rate of international high-precision MEMS gyroscopes is 0.01°/h, while domestic ones are about 0.1°/h; the ranging distance of international automotive-grade LiDAR can reach 200m, while mainstream domestic products are within 180m; the frame rate of international industrial-grade vision reaches 1000fps, while domestic ones are mostly below 500fps. In addition, the long-term stability life of international products is 5-10 years, while domestic ones are 3-5 years. Gaps still exist in extreme environment adaptability, such as high temperature and strong corrosion;
(3) Weak basic research: Insufficient investment in basic research fields such as sensing materials (e.g., piezoelectric materials, magnetoresistive materials), MEMS processes (e.g., deep silicon etching, bonding technology), and detection algorithms (e.g., high-precision calibration algorithms, AI defect recognition algorithms). Patent layout is concentrated at the application layer, with basic core technology patents accounting for less than 10%, which has been monopolized by European and American enterprises for a long time.
2. Integration Difficulties: Interconnection Barriers
(1) Inconsistent protocols: There are multiple communication protocols in the current industrial sensor market, such as PROFINET, EtherCAT, Modbus, LoRa, and NB-IoT. The interfaces and communication protocols of products from different manufacturers vary greatly, with insufficient compatibility. Sensors from Factory A support the PROFINET protocol, while PLCs from Factory B support the EtherCAT protocol. The two need to convert data through gateways to realize data interaction, increasing system complexity and costs;
(2) Inefficient collaboration of heterogeneous equipment: The effective data interaction rate between sensors and PLCs, MES systems, and digital twin platforms is less than 70%, forming "data silos." On the one hand, sensor data formats are not unified (e.g., temperature data has two units: ℃ and ℉, and pressure data has two units: MPa and PSI), requiring additional data conversion; on the other hand, the data transmission rates and storage formats of different systems vary greatly, leading to data synchronization delays (in some scenarios, the delay reaches over 100ms), restricting full-process digital collaboration;
(3) Data security risks: Scenario data contains core information such as production processes, equipment parameters, and product quality, which is vulnerable to eavesdropping, tampering, and forgery attacks in wireless communication scenarios. However, low-power sensors are limited by computing power and power consumption, making it difficult to deploy complex security protection mechanisms (such as high-strength encryption algorithms), resulting in a high risk of data leakage. For example, a wireless sensor network in an automotive factory once suffered a malicious attack, leading to the leakage of production process parameters and causing economic losses of over 10 million RMB.
3. Cost Pressure: High Application Threshold for SMEs
(1) High R&D costs: Sensors involve multi-disciplinary integrated technologies, with long R&D cycles and large investments. A single MEMS chip production line requires an investment of over 1 billion RMB, and a packaging and testing production line requires an investment of over 200 million RMB; the proportion of R&D investment in applications alone reaches 18%, far higher than the 5%-8% in ordinary manufacturing. SMEs have limited financial strength and find it difficult to bear continuous R&D investments, leading to weak technological innovation capabilities;
(2) High procurement and maintenance costs: Industrial discrete sensors are expensive. For example, the unit price of imported automotive-grade LiDAR is about 15,000 RMB, while domestic industrial-grade ones are about 5,000-8,000 RMB; the unit price of imported vision sensors is about 8,000 RMB, while domestic ones are about 2,000 RMB. In addition, sensors require regular calibration and maintenance, and maintenance costs account for over 20% of the total equipment cost, which is hard for SMEs to bear, restricting the large-scale application of sensors;
(3) Insufficient economies of scale: There are numerous sensor enterprises, but they are small in scale. Less than 20 enterprises have annual revenues exceeding 1 billion RMB, and market concentration is low (CR10 < 30%). Enterprises lack economies of scale, leading to high production costs and making it difficult to compete with international giants. For example, the unit cost of mass production by international MEMS chip giants can be reduced to less than 1 US dollar, while the unit cost for SMEs in China is about 3-5 US dollars.
Three Breakthrough Directions
1. Localization Development: Focusing on Core Link Research
(1) Component breakthroughs: Concentrate efforts to overcome core links such as MEMS chips, ASIC circuits, and sensing materials to increase the domestic production rate of chips; encourage enterprises to develop dedicated chips and ASIC circuits for industrial sensors matching industrial-grade applications; break through sensing material technologies, such as piezoelectric materials (PZT), magnetoresistive materials (GMR/TMR), and gas sensing materials, to reduce dependence on imported materials.
(2) Technological research: Establish a collaborative innovation mechanism of "industry-university-research-application," concentrate efforts to break through high-precision detection algorithms, high-reliability packaging technologies, and extreme environment adaptability technologies. Support enterprises in participating in the formulation of international standards to enhance technological discourse power;
(3) Policy support: Relying on policies such as the "Guiding Catalogue for Industrial Structure Adjustment" and "Guiding Opinions on Accelerating the Promotion of Intelligent Transformation in the Industrial Field," increase investment in R&D subsidies and pilot test platform construction. Set up national and local industrial special funds to support enterprise mergers, reorganizations, and technological upgrades; strive to break through 65% in the domestic production rate of industrial high-end sensors and increase the domestic production rate of core components to over 50% by 2030.
2. Standardization Construction: Breaking Through Collaboration Barriers
(1) Unified interfaces and protocols: Led by the Ministry of Industry and Information Technology and the Standardization Administration of China, formulate national standards for industrial sensor data interfaces and communication protocols. Promote the domestic adaptation of international mainstream protocols such as PROFINET and EtherCAT, formulate unified data format standards, and realize cross-manufacturer and cross-system compatibility. For example, the "Unified Specification for Industrial Sensor Communication Protocols" released by the China Sensor and IoT Industry Alliance has been piloted in industries such as automotive manufacturing and electronic information, increasing the data interaction rate to over 90%;
(2) Establish a testing and certification system: Set up a national-level sensor testing and certification center to standardize product consistency testing. For example, formulate testing standards for LiDAR ranging accuracy, response time, and environmental adaptability, and ensure product quality through third-party testing and certification; establish a sensor traceability system to ensure the accuracy and reliability of testing data and reduce system integration risks;
(3) Data security protection: Follow international standards to build a full life cycle security protection system for sensor data. Hardware can use security chips to realize identity authentication and data encryption; at the software level, technologies such as end-to-end encryption, message authentication code (MAC) verification, and access control are adopted to prevent data leakage and tampering.
3. Ecological Cooperation: Lowering Application Thresholds
(1) Industry-university-research-application collaboration: Universities, research institutes, and enterprises jointly build pilot test platforms and joint laboratories, drive innovation through scenarios, accelerate achievement transformation, and strengthen R&D innovation momentum.
(2) Scenario-based solutions: For subdivided fields such as NEVs, semiconductors, and biomedicine, provide an integrated solution of "sensors + data platform + operation and maintenance services," offering full-process services including sensor deployment, data collection, analysis and warning, and operation and maintenance services, reducing deployment costs for SMEs.
(3) Industrial clusters: Through the construction of regional and park ecological systems, create advantageous regions, such as multiple hundred-billion-level sensor industrial clusters or "Sensor Valleys" in the Yangtze River Delta, Guangdong-Hong Kong-Macao Greater Bay Area, and Beijing-Tianjin-Hebei, realizing resource sharing in chip design, packaging and testing, and terminal applications upstream and downstream.
(4) Cost control: Reduce costs through mass production, process optimization, and supply chain integration. Sign long-term cooperation agreements with sensing material and chip manufacturing enterprises to reduce raw material procurement costs.
Future Trends and Industrial Outlook
Given the iterative upgrading of industrial intelligent scenarios and the continuous expansion of demand, sensor technology will evolve towards miniaturization, low power consumption, intelligence, and multi-dimensional perception. The industry will transform from a discrete, multi-variety, and small-batch production model to industrialized scale production.
Technological Trends: Breakthroughs in Four Core Directions
1. Miniaturization and Integration: The development of MEMS technology and nanotechnology will drive the continuous shrinking of sensor sizes, breaking through from the millimeter level to the micrometer and nanometer levels. For example, the size of MEMS chips will achieve ultra-miniaturization of 0.1mm×0.1mm×0.05mm in the future; "optical-acoustic-force" integrated sensing devices will become mainstream. By integrating various sensing units, microprocessors, and communication modules, the integration of "perception-computation-communication" is realized, meeting the deployment needs of miniaturized equipment (such as micro-robots and wearable devices);
2. Low Power Consumption and Long Life: Optimize circuit design (such as using low-power MCUs and sleep technology) and packaging technology (such as vacuum packaging and low-power materials) to reduce power consumption to below the milliwatt level (currently 10-100 milliwatts) and extend the service life to over 10 years, adapting to large-scale and long-cycle deployment scenarios of the Industrial IoT; for example, using micro-energy harvesting technology (such as electromagnetic induction, vibration, temperature difference, and other industrial scenario-adapted micro-energy harvesting technologies) can realize self-powering without the need to replace batteries, further reducing maintenance costs;
3. Self-calibration and Intelligence: Built-in AI self-calibration algorithms automatically adjust parameters according to environmental changes (such as temperature and humidity) to realize environmental adaptive adjustment and reduce manual maintenance costs; upgrading from "data acquisition" to "intelligent cognition," sensors have autonomous decision-making and collaborative working capabilities. For example, multi-sensor collaborative perception systems realize data fusion and decision-making through AI algorithms, can independently choose the optimal perception strategy, and adapt to complex scenario needs;
4. Multi-dimensional Perception: Build an array-based and networked full-domain perception system. For example, large-scale sensor arrays (100,000+ sensing points) can realize large-area and high-precision perception; through IoT technology, a "device-edge-cloud" collaborative perception system is formed to support multi-parameter and high-precision monitoring in complex scenarios; in addition, breakthroughs in emerging sensing technologies (such as quantum sensing and flexible sensing) will expand perception dimensions, realizing precise perception of special physical quantities such as weak magnetic fields and biological signals.
Application Expansion: Extension from Traditional Manufacturing to Emerging Fields
1. Traditional Manufacturing Upgrading: Sensors will deeply penetrate traditional manufacturing industries such as automotive, electronics, machinery, and chemical engineering, realizing full-element perception and closed-loop control of production processes. For example, the "perception-decision-execution" closed-loop control in the manufacturing processes of automobiles and home appliances will deploy massive sensors to realize full-process intelligent and unmanned functions; electronic manufacturing will realize full-process automated inspection of chip manufacturing, packaging, and testing, with sensor accuracy improved to the nanometer level, promoting the development of electronic products towards higher performance and smaller sizes; it is expected that by 2030, the application coverage rate of sensors in traditional manufacturing will reach over 90%, and production efficiency will increase by over 20%;
2. Emerging Field Expansion, Presenting New Growth Points: The digitalization and intelligence of industrial processes in emerging fields such as NEVs, semiconductors, biomedicine, humanoid robots, and intelligent construction will give rise to new application scenarios. It is expected that by 2030, the market size of sensors in the digitalization field of manufacturing will break through 360 billion RMB.
Conclusion
As the basic core engine of intelligent manufacturing, the technological innovation and industrial development of sensors directly determine the depth and effectiveness of the digital transformation of the manufacturing industry. From the precise and efficient front-end material distribution, to the precise control of the mid-end manufacturing process, and then to the intelligent optimization of back-end equipment operation and maintenance, sensors have permeated every link of intelligent manufacturing, building a closed-loop system of "comprehensive perception, intelligent decision-making, and efficient execution."
Currently, although China's sensor industry faces challenges such as high-end technological bottlenecks, integration barriers, and cost pressure, driven by multiple factors such as policy support, technological research, and ecological collaboration, domestic sensors are accelerating the breakthrough of core technologies, expanding application scenarios, and improving the industrial ecology.
In the future, with the continuous evolution of miniaturization, low power consumption, intelligence, and multi-dimensional perception technologies, sensors will further break through performance limits and application boundaries, playing a greater role in traditional manufacturing upgrading and emerging field expansion. It will promote the leap of the manufacturing industry from "automation" to "intelligence," inject continuous power into industrial digitalization and high-quality development, and help write a new chapter in the construction of a manufacturing powerhouse.
(The next issue of this column will feature the "Smart Agriculture Sensors" article, stay tuned!)
Author | Guo Yuansheng, Deputy Director of the Science and Technology Committee of the Central Committee of the Jiusan Society, Executive Vice Chairman of the China Sensor and IoT Industry Alliance
Editor | Yang Pengyue Art Editor | Ma Liya Supervisor | Zhao Chen