How many sensors does a robot need to be as sensitive as a human?
AsiaIndustrial Netnews: today’srobotIt already has human-like limbs and sensory functions, a certain degree of intelligence, flexible action programs, and can work without relying on human manipulation. And all of this is due to the sensor. The sensor is an important helper for the robot to perceive the outside world. They are like human perception organs. The robot’s vision, force, touch, smell, taste and other perceptions of the external environment are all provided by the sensor. At the same time, the sensor can also be used to detect the working state of the robot itself, and the robot can intelligently detect the external working environment and the state of the object. A device that can be converted into a usable output signal according to certain rules. In order for the robot to achieve the highest possible sensitivity, various sensors will be installed in its body structure. So how many sensors should the robot have? Can it be as sensitive as human beings as possible?
Robot sensors can be divided into internal sensors and external sensors according to the detection object.
Internal sensors are mainly used to detect the status of each internal system of the robot, such as the position, speed, acceleration temperature,motorSpeed, motor load, battery voltage, etc., and the measured information is sent to the controller as feedback information to form a closed-loop control.
The external sensor is used to obtain information about the robot’s work object and the external environment. It is an information channel for the robot to interact with the surroundings. , light, etc.
The details are as follows:
1. Vision sensor
Machine vision is a system that enables Robots to have perception functions. It acquires images through visual sensors for analysis, allowing robots to replace human eyes in identifying objects, measuring and judging, and realizing positioning and other functions. People in the industry pointed out that at present, the smart vision sensors that are easy to use in China account for about 60% of the market share of the machine vision system market.The advantage of the visual sensor is that it has a wide range of detection and rich information.In the application, multiple vision sensors are often used or used in conjunction with other sensors. Through a certain algorithm, many information such as the shape, distance, and speed of the object can be obtained.
The field of computational vision based on depth cameras has become one of the hottest investment and entrepreneurial hotspots in the entire high-tech industry. Interestingly, many cutting-edge achievements in this field were first launched by start-up companies and then carried forward by giant acquisitions. For example, Intel acquired the RealSense real-sensing camera, Apple acquired PrimeSense, a Kinect technology supplier, and Oculus acquired a company that focuses on high-precision Pebbles Interfaces, an Israeli technology company with gesture recognition technology. Although the entrepreneurial teams in domestic computing vision have not entered the mainstream vision of investors on a large scale, the best ones have begun to achieve remarkable results.
The concept of depth camera was proposed by IBM as early as the 1980s. This super company, which holds almost all the underlying data of the hard disk in the past, present and future, can be described as the leader of the times. Founded in Israel in 2005, PrimeSense is a pioneer in the civilian use of this technology.At that time, the promotion of depth cameras in the consumer market was still in the conceptual stage. Previously, depth cameras were only used inindustryfield formechanicalarm,Industrial robots, etc. provide graphics and vision services. Microsoft Kinect, which provides technical solutions, has become the pioneering work of depth cameras in the consumer field, and has driven the civilian development of this technology throughout the industry.
1 2 3 Next > page
2. Acoustic sensor
The sound sensor acts like a microphone (microphone). It is used to receive sound waves and Display the vibration image of the sound. However, the intensity of the noise cannot be measured. Acoustic sensors are mainly used to sense and interpret sound waves in gases (non-contact sensing), liquids or solids (contact sensing). The sophistication of acoustic sensors can range from simple acoustic presence detection to complex acoustic frequency analysis to the discrimination of individual sounds and words in continuous natural language.
It is reported that since the 1950s, Bell Labs has developed the world’s first speech recognition Audry system, which can recognize 10 English numbers. In the 1970s, the sound recognition technology developed rapidly, and the Dynamic Time Warping (DTW) algorithm, Vector Quantization (VQ) and Hidden Markov Model (HMM) theories were proposed one after another, and the isolated speech of a specific person based on DTW technology was realized. recognition system. In recent years, voice recognition technology has moved from the laboratory to practical use, and many companies at home and abroad have developed corresponding products using voice recognition technology. More well-known companies include Speech, HKUST Xunfei, Tencent, Baidu and other giants, and they have entered the field of voice technology.
3. Distance sensor
The distance sensors used for intelligent mobile robots include laser rangefinders (also capable of measuring angles), sonar sensors, etc. The lidar sensor developed in recent years is currently the mainstream one, which can be used for robot navigation and obstacle avoidance. For example, the RPLIDAR A2 lidar developed by SLAMTEC can perform 360-degree scanning and ranging to obtain the contour map of the surrounding environment. The sampling frequency is as high as 4000 times per second, which has become the highest measurement frequency of low-cost lidar in the industry. Cooperating with SLAMTEC’s SLAMWARE autonomous positioning and navigation solution, it can help robots realize autonomous map building, real-time road planning and automatic obstacle avoidance.
4. Tactile sensor
Tactile sensors are mainly sensors used in robots to mimic tactile functions. The sense of touch is an important sensory function when human beings are in direct contact with the external environment, and the development of tactile sensors that meet the requirements is one of the key technologies in the development of robots. With the development of microelectronics technology and the emergence of various organic materials, a variety of development schemes for tactile sensors have been proposed, but most of them belong to the laboratory stage at present, and few have reached commercialization.
5. Proximity sensor
The proximity sensor is between the tactile sensor and the visual sensor, which can measure the distance and orientation, and can fuse the information of the visual and tactile sensors. The proximity sensor can assist the function of the visual system to judge the orientation and shape of the object, and recognize its surface shape at the same time. Therefore, in order to accurately grasp the parts, the precision requirements for the proximity sensor of the robot are very high. This sensor mainly has the following functions:
Find obstacles ahead and limit the range of motion of the robot to avoid collisions with obstacles.
Get the necessary information before touching the object, such as the relative distance and relative inclination to the object, so as to prepare for the subsequent actions. Obtain the distance between points on the surface of the object, so as to obtain information about the shape of the object’s surface.
page
6. Slip sensor
The slip sensor is mainly a sensor used to detect the degree of slippage between the robot and the grasped object. In order to determine an appropriate grip force value when grasping an object, it is necessary to detect the relative sliding of the contact surface in real time, then judge the grip force, and gradually increase the force without damaging the object. The slip detection function is necessary to realize the flexible grip of the robot. condition. The recognition function can be realized through the slip sensor, and the surface roughness and hardness of the grasped object can be judged. Slip sensors can be divided into three categories according to the sliding direction of the measured object: non-directional, unidirectional and omnidirectional sensors. Among them, the non-directional sensor can only detect whether there is sliding, but cannot distinguish the direction; the unidirectional sensor can only detect the sliding in one direction; the omnidirectional sensor can detect the sliding in one direction. This sensor is generally made into a spherical shape to meet the needs.
7. Force sensor
The force sensor is a sensor used to detect the interaction force between the robot’s own force and the external environment force. Force sensors are often installed at the joints of the robot, and indirectly measure the force by detecting the deformation of the elastic body. The force sensor installed in the joint of the robot often appears in a fixed three-coordinate form, which is beneficial to meet the requirements of the control system. The current six-dimensional force sensor can realize the measurement of full force information, and it is called wrist force sensor because it is mainly installed at the wrist joint. Most of the wrist force sensors adopt the principle of electric strain measurement, and can be divided into two types according to their elastic body structure, cylindrical and cross-shaped wrist force sensors. Among them, the cylinder type has the characteristics of simple structure, high utilization rate of the elastic beam, and high sensitivity; while the cross-shaped sensor has a simple structure, easy coordinate establishment, but high processing accuracy.
8. Speed and acceleration sensor
There are two types of speed sensors for measuring translational and rotational speeds, but in most cases, they are limited to measuring rotational speeds. Using the derivative of the displacement, especially the photoelectric method, let the light irradiate the rotating disk, detect the rotation frequency and the number of pulses, to find the rotation angle, and use the disk to make a gap, and distinguish the angular velocity through two photodiodes, that is, the rotational speed , This is the photoelectric pulse speed sensor.
An accelerometer is a sensor capable of measuring acceleration. It usually consists of mass block, damper, elastic element, sensitive element and adjustment circuit. During the acceleration process, the sensor obtains the acceleration value by using Newton’s second law by measuring the inertial force on the mass block. According to the different sensitive elements of the sensor, common acceleration sensors include capacitive, inductive, strain, piezoresistive, piezoelectric, etc.
In order for a robot to be as sensitive as a human being, the eight sensors of visual sensor, acoustic sensor, distance sensor, tactile sensor, proximity sensor, force sensor, slip sensor, speed and acceleration sensor are extremely important to the robot, especially The five sensory sensors of the robot are essential. From the perspective of anthropomorphic functions, vision, force, and touch are the most important, and they have entered the practical stage, but its senses, such as hearing, smell, taste, sliding, etc. The sensors are still waiting to be conquered one by one.
The Links: 3BSE042238R1 SH-4
Pre: Vicor launches ‘Power-Driven In... Next: How should chemical companies improve...