With the development of IoT technology, new requirements have been put forward for traditional sensing technology. Products are gradually developing towards MEMS technology, wireless data transmission technology, infrared technology, new materials technology, nanotechnology, ceramic technology, thin film technology, fiber optic technology, laser technology, composite sensor technology, and multidisciplinary integration.
Sensors are crucial hardware components for devices to sense their external environment, determining their ability to interact with it. They form the hardware foundation for device intelligence, especially in many smart devices where sensors define the core capabilities. A typical sensor consists of a sensing element, a conversion element, and a conditioning circuit. The sensing element directly senses the measured quantity, while the conversion element transforms it into an electrical parameter.
Sensors have a wide range of applications and come in many types. Based on the object being measured, sensors can be categorized into those that detect light, radiation, sound signals, magnetic signals, force, position information, temperature, humidity, and solution flow rate/velocity. Each type of sensor detecting the same object has multiple applications and different implementation paths.
Currently, the focus of domestic sensor technology development and innovation is on three aspects: materials, structure, and performance improvement. Sensitive materials are developing from liquid to semi-solid and solid states; structures are developing towards miniaturization, integration, modularization, and intelligence; and performance is developing towards wider detection range, higher detection accuracy, stronger anti-interference ability, stable performance, and longer lifespan.
With the development of IoT technology, new requirements have been put forward for traditional sensing technology. Products are gradually developing towards MEMS technology, wireless data transmission technology, infrared technology, new materials technology, nanotechnology, ceramic technology, thin film technology, fiber optic technology, laser technology, composite sensor technology, and multidisciplinary integration.
Smart sensors give intelligent equipment a variety of "senses".
Sensors, as the only autonomous input for intelligent equipment, are equivalent to the various sensory organs of intelligent equipment and robots. The main senses that intelligent equipment has for the external environment include vision, position sense, speed sense, force sense, and touch sense.
Vision is the most commonly used input system for intelligent equipment and can be divided into two main categories: one is intuitive vision, where the data type is an image composed of pixels, and typical applications include machine vision and object recognition. Sensors of this type include high-speed cameras and video cameras; the other is environmental model-based vision, where the data type is a spatial model composed of point cloud data, and typical applications include spatial modeling. Sensors of this type include 3D LiDAR and laser scanners.
Position sense refers to the ability to determine one's location in an environment by sensing the distance between oneself and surrounding objects. Such sensors include laser rangefinders, 2D lidar, magnetometers (for determining direction), millimeter-wave radar, and ultrasonic sensors.
Speed sensing refers to the ability of intelligent equipment to perceive its own speed, acceleration, angular velocity, and other information. Such sensors include speed encoders, accelerometers, and gyroscopes.
Force sensing is used in intelligent equipment to detect forces from external contact objects or internal mechanical mechanisms. Typical applications include force sensors mounted on joint actuators to provide force feedback, and force sensors mounted between the end effector of a robotic arm and the last joint of a robot to detect forces applied by objects.
In intelligent equipment, tactile sensation can be further divided into contact sensation, pressure sensation, and slip sensation. Such sensors include optical tactile sensors, piezoresistive array tactile sensors, and slip sensors. Among them, slip sensors are essential for realizing the grasping function of robots.
In addition to the five human senses mentioned above, some physical sensors possess capabilities that surpass those of the human senses. For example, biosensors can measure blood pressure and body temperature, while environmental sensors can measure temperature and humidity, airborne particulate matter levels, and ultraviolet light intensity. These sensors, which extend human senses, are now being used in wearable devices, thus giving these devices the ability to expand upon human sensory capabilities.
Sensor R&D Trends
With the advancement of technologies such as microelectromechanical systems (MEMS), laser technology, and high-tech materials, sensor research and development is showing a diversified trend. Some sensors use biomaterials to simulate human skin and innovate the tactile sensation of sensors; some use MEMS technology to develop miniature intelligent sensors, which is conducive to the integration of complex systems; and some use high-precision laser technology to create lidar, which is conducive to the system's real-time perception of surrounding obstacles and environment, etc.
However, in general, the sensor R&D process exhibits a two-stage trend: First, technological innovation, developing new products based on unmet needs. In the first stage, sensor R&D innovation stems from the demands of intelligent equipment and innovative devices; researchers innovate new sensors based on usage requirements. Second, cost reduction and application implementation, with products gradually meeting industrialization needs. In the second stage, during the R&D innovation process, to meet the demands for the industrial application of intelligent equipment, researchers shift their focus from technological development to cost reduction, aiming to achieve the vision of large-scale sensor production and the industrial application of intelligent equipment.
3D LiDAR is a type of sensor that originated from functional innovation and has now entered the stage of commercial development. The following section uses LiDAR as an example to outline typical development trends in sensors.
Research and Development Trend 1:
The emergence of 3D LiDAR, a technology that extends to higher levels, was created to meet the system's need for real-time spatial perception. Autonomous mobile robots such as self-driving cars and drones require a sensor that can scan the surrounding environment in real time to obtain information about the distance to obstacles and roads, in order to achieve spatial recognition, autonomous obstacle avoidance, and route planning. Thus, 3D LiDAR came into being.
The development of 3D LiDAR is essentially an upgrade of laser ranging technology and a process of gradually upgrading the needs it meets. Laser ranging technology is the foundation of 3D LiDAR. The earliest laser rangefinders solved the need for point-to-point one-dimensional distance measurement; then 2D LiDAR solved the need to perceive approaching objects within a fan-shaped plane, measuring distances within that plane; now, 3D LiDAR, through high-speed changes in the laser projection angle, scans the surrounding environment in real time to obtain distance information, solving the need for obstacle and environmental recognition in three-dimensional space, measuring distances within three-dimensional space.
The most popular application of 3D LiDAR is undoubtedly in autonomous vehicles. 3D LiDAR-based autonomous driving perception systems represent the mainstream technology in the field today. However, the cost of 3D LiDAR has always been a major drawback. Taking Velodyne, a leading manufacturer of 3D LiDAR, as an example, their three products, priced from highest to lowest performance, are $80,000, $20,000, and $8,000 respectively. During the research and testing phase of autonomous vehicles, research institutions including Google and Baidu have consistently used the $80,000 version for testing. It is understood that the total cost of Google's autonomous vehicle is approximately $300,000, and the 64-line HDL-64 3D LiDAR accounts for 25% of the total vehicle cost.
Research and development trend two:
After the first stage of technological innovation in R&D, the high cost became the main problem for autonomous driving perception systems, primarily based on 3D LiDAR. Sensor manufacturers shifted their focus from enhancing functionality to cost control in LiDAR R&D, thus entering the second stage of R&D: reducing costs to achieve industrial application.
At CES 2016, often referred to as a "technology trendsetter in the consumer electronics sector," LiDAR technology companies Velodyne and Quanergy both showcased new 3D LiDAR systems. Velodyne's PuckAuto and Quanergy's S3 are both miniaturized improvements over their predecessors.
Velodyne's PuckAuto uses a 32-line laser with a scanning range of 200 meters. It can be considered an enhanced version of the VLP-16, better suited to the needs of autonomous vehicles, and more cost-effective than the other two products. The company has reached a cooperation agreement with Ford, and Ford's Fusion autonomous vehicles will be equipped with two PuckAuto units. Velodyne executives stated they will further reduce product costs, aiming to keep the price below $1,000.
Quanergy's S3 is a solid-state LiDAR developed in collaboration with Delphi. It uses an 8-line laser, has no internal rotating parts, and can be integrated into the vehicle. Previous reports indicated that Quanergy's CTO stated each S3 costs around $200. This extremely low price is due to the product's configuration; the "8-line" and "solid-state" characteristics allow for effective cost control. "Solid-state" means it cannot rotate 360 degrees and can only detect what is in front, but this limited detection range can be compensated for by quantity. Placing four or six S3s at the four corners of the vehicle is a solution explored by Delphi's autonomous vehicles.
Through the new products launched by these two American technology companies at CES 2016, we can see that the technical characteristics of LiDAR are gradually meeting the industrialization needs of the autonomous driving field. After removing redundant hardware configurations in the testing phase, the cost is expected to be significantly reduced.
Sensor application trends: Combining similar technologies, using multiple technologies, and innovating application scenarios
As the only input to intelligent equipment besides manually set parameters, the importance of sensors is self-evident. The ability of sensors to perceive the external environment determines the accuracy and richness of the information input to intelligent equipment. Innovations in the effective application of sensors often form the basis for functional innovations in intelligent equipment. The innovative application of sensors in intelligent equipment mainly follows three trends:
Using similar sensors in combination, achieving vertical integration of single functions.
In such cases, the system often has extremely high requirements for a single function. To meet the highly complex requirements of the system in a single function, similar sensors are organically combined to form a redundant structure that ensures the system's safety in that function. For example, in the perception system of an autonomous vehicle, the organic combination of multiple visual and positional sensors forms a complementary redundant structure, thereby ensuring that the system can correctly and efficiently perceive the external environment in real time and make correct driving decisions.
At this point, sensors often have a dominant and auxiliary function, and the dominant sensor is the core technological barrier for product realization.
The combination of multiple sensors and the wide range of functions can be combined in a horizontal way.
To meet the diverse and multi-layered input and output needs of the system, various types of sensors are innovatively combined to form multiple senses in intelligent equipment, generating intelligent feedback based on these senses. For example, the emotionally interactive robot Pepper, as well as other companion and early education robots, combine multiple senses to form an emotional perception system encompassing vision, position sense, and hearing, which then generates intelligent feedback through internal artificial intelligence algorithms.
At this point, there is no distinction between primary and secondary hardware components; both system and algorithm chips play equally important roles.
New sensors are applied to traditional equipment, giving the equipment a smart vitality.
New intelligent sensors are being applied to traditional devices, giving them "senses" and thus upgrading them into intelligent devices. For example, the combination of LiDAR and robotic vacuum cleaners creates path-planning robotic vacuum cleaners; the combination of blood pressure sensors, heart rate sensors, and position sensors with watches and wristbands creates wearable devices that integrate various health monitoring functions.
In this situation, since there is demand for traditional equipment, it is mainly a phenomenon of penetration and replacement in the existing market, while the improved performance brought about by the application of new sensors has a clear consumer base.
For more information, please follow the Sensor Channel.
