The six-dimensional torque sensor can control the optimal solution, and the technical barrier is high

KJ3202X1-BA1 According to the measurement direction, the six-dimensional torque sensor is the one with the best performance and the most comprehensive force perception information. In a specified rectangular coordinate system, the six-dimensional torque sensor can simultaneously measure forces along three axes (F, FY, FZ) and moments around three axes (MX, MY, MZ). The six-axis force sensor is generally divided into a fixed end (robot end) and a loading end (tool end). When the two ends are relative to each other, the sensor will undergo elastic deformation, and the resistance of the strain gauge inside the sensor will change, which will be converted into a voltage signal output.

In the requirement of accurate measurement, the six-dimensional torque sensor is the best choice. If the direction of the force and the point of application are fixed, a one-dimensional force sensor can be selected; If the direction of the force changes randomly, but the application point remains unchanged and coincides with the calibration reference point of the sensor, a three-dimensional force sensor can be selected. If the direction of the force and the point of application are random changes in the three-dimensional space, a six-dimensional force sensor should be used for measurement. The high-precision six-dimensional torque sensor can decouple the interference between the forces in all directions and the torque, so that the force measurement is more accurate, and the torque information can be used to calculate the attitude of the stressed parts, monitor whether the torque is within the safe range, and effectively avoid the overload damage of the sensor.

KJ3202X1-BA1 Force sensors are widely used and humanoid robots are widely configured. Force sensors are widely used in industry, aerospace, automotive, medical equipment and other fields. In the field of robotics, force sensors can be used to measure the force on the robot joints in real time, and realize the main power output control, which plays an important role in high-complexity work and coordination work. The force sensor in the robot is mainly a single-axis torque sensor related to the node position and a six-axis force sensor at the end of the robot actuator to measure the internal force of the robot and the force interaction between the end actuator and the external environment. The six-dimensional torque sensor has a variety of industrial application scenarios, mainly including grinding, precision assembly, medical treatment, special operations, testing and other scenarios involving contact operation and requiring multi-dimensional force perception, among which the robot field is more applied, and is the core application field of six-dimensional force/torque sensor.

The high precision control of robot motion brings the demand of six-dimensional torque sensor. The six-dimensional force sensor on the end joint of the collaborative robot usually needs to be connected with the actuator such as the grinding head and the clamp. The moment arm of the actuator varies from tens of mm to 300 mm during the working process, and the moment arm is large and random, so the six-dimensional force sensor is usually used to achieve high-precision control. In humanoid robots, force/moment sensors are usually installed between the sole of the foot and the ankle joint and between the manipulator and the wrist joint to provide more comprehensive force perception.

KJ3202X1-BA1

Multi-dimensional force sensor miniaturization is an important development trend in the future. In order to take into account the miniaturization and high performance of multi-dimensional force sensors, the main way that can be taken at present is to choose a small size (or volume) of sensitive elements, so as to reduce the size of the sensor structure. Piezoresistive sensitive elements (such as silicon strain gauges, etc.) mostly use semiconductor materials and processing technologies, which have significant advantages in volume, performance and cost, and have become one of the important technical ways to achieve the miniaturization of multi-dimensional force sensors. The application of semiconductor silicon strain gauges to achieve miniaturization of multi-dimensional force sensors involves specific technical problems: the design of silicon strain gauges and its mounting process need to meet the design and use requirements KJ3202X1-BA1 of multi-dimensional force sensors. At present, the mainstream commercial pressure sensors mostly use (integrated) silicon strain gauges. The traditional manufacturing process of silicon strain gauges in China is to doping N-type impurities on P-type silicon wafer substrate by diffusion or ion implantation to form varistor, and to form strain gauges by photolithography, corrosion, lead bonding, etc., but it is easy to cause large resistance difference between strain gauges. The strain calculation used by multi-dimensional force sensors has a higher demand for consistency, and poor consistency will lead to a large difference in the zero output of different Bridges (component forces), which is not conducive to its subsequent temperature and decoupling signal processing and compensation, and hinder the mass production of sensors. For the patch process, traditional force sensors usually use organic adhesives (silicone, epoxy, etc.) to bond metal (or semiconductor) strain gauges to the sensor elastic elements, which is the most common sensor patch material at present.

The technical barriers of six-dimensional torque sensor are extremely high, covering many aspects. In addition to the variety of forms, the development of six-dimensional force sensors is also difficult, and its nonlinear mechanical characteristics are obvious. It is necessary to consider the temperature drift, creep, cross interference of multi-channel signals, real-time data processing, loading and calibration complexity, etc., and the technical barriers are mainly in the structural decoupling design, decoupling algorithm, calibration and calibration.

Structural decoupling design is an important barrier. The form and arrangement of the force sensing element directly affect the sensitivity, stiffness, dynamic performance and dimensional coupling of the sensor, and determine the performance of the sensor to a large extent. Common structural designs include vertical beams, beams and other integrated structures as well as Stewart parallel platforms. The vertical beam structure has good lateral effect, simple structure and strong bearing capacity, but poor vertical effect, large interdimensional interference and low sensitivity. The cross beam structure has high sensitivity, easy machining and calibration, but there are dimensional coupling and radial effects. The elastomer used in the Stewart platform is a composite structure. This type of sensor has the advantages of compact structure, strong load bearing capacity and non-accumulation of errors. Patch position adjustment and bridge group bridge design further eliminate coupling. The special structural design can greatly reduce the coupling deformation of the force sensing element, but the structural continuity leads to the coupling deformation of the structure can not be completely avoided, so it is necessary to further eliminate the coupling patch by mounting and bridge assembly.

The decoupling algorithm will decouple the interference between the forces and moments in each direction, so that the force measurement is more accurate. The input signal in each of the six channels of the six-component force sensor will affect the output signal of the other channels, and KJ3202X1-BA1 decoupling is to reduce or eliminate the coupling interference to a certain extent. Generally, coupling elimination or coupling suppression can be done from two directions, the first is the work carried out before the production of sensors, generally called structural decoupling, that is, from the design of the sensor to eliminate or inhibit coupling; The second is to use a systematic mathematical model, to matrix decoupling, the use of digital signal processing method to reduce or eliminate the sensor’s dimensional coupling. This method has low requirements for manufacturing, is relatively easy to achieve, and can also achieve good results. Calibration and inspection calibrate the sensor accuracy. By loading the theoretical load and recording the corresponding original signal output of the sensor, the mapping relationship between the original signal and the force of the sensor is established, and the mathematical model and parameters of the decoupling algorithm are obtained. After the calibration is completed, the sensor accuracy and accuracy can be obtained by loading the load known to the theoretical truth value and recording the measurement results of the sensor at the same time. Simply put, calibration is to obtain sensor firmware parameters, and detection is to obtain sensor accuracy.