Robot Joint Torque Research Articles

Joint torque estimation method of industrial robot by combining error compensation model based on LSTM and iterative weighted parameter identification

It is essential for human-robot interaction to accurate joint torque value estimation of industrial robots without force/torque (F/T) sensors. The most common method to estimate the value of robot joint torque is based on dynamic model parameter identification. However, no matter how elaborate the modeling methods are used, there are always uncertainty errors when robot’s joint rotation direction changes. In this paper, an identification framework is proposed to estimate optimal model parameters, and a deep learning network is proposed to compensate for the uncertainty errors caused by the unmodeled dynamics part. At first, the dynamic parameters are estimated during the Least Squares (LS) identification process considering the noise influence and data outliers, and the nonlinear friction model is unified into the identification framework. Secondly, based on Long-Short Term Memory (LSTM) network, an error compensation model (ECM) is proposed to establish the mapping relationship between the joint motion and the identification errors. Finally, a 6-DOF robot is used for parameter identification and ECM validation. Experimental results show that the proposed identification method is better than the LS method. Compared with the torque value estimated by the identification method, the proposed ECM can compensate for the uncertainty errors and improve the torque estimation accuracy.

Strain Gauge Neural Network-Based Estimation as an Alternative for Force and Torque Sensor Measurements in Robot Manipulators

When a robotic manipulator interacts with its environment, the end-effector forces need to be measured to assess if a task has been completed successfully and for safety reasons. Traditionally, these forces are either measured directly by a 6-dimensional (6D) force–torque sensor (mounted on a robot’s wrist) or by estimation methods based on observers, which require knowledge of the robot’s exact model. Contrary to this, the proposed approach is based on using an array of low-cost 1-dimensional (1D) strain gauge sensors mounted beneath the robot’s base in conjunction with time series neural networks, to estimate both the end-effector 3-dimensional (3D) interaction forces as well as robot joint torques. The method does not require knowledge of robot dynamics. For comparison reasons, the same approach was used but with 6D force sensor measurements mounted beneath the robot’s base. The trained networks showed reasonably good performance, using the long-short term memory (LSTM) architecture, with a root mean squared error (RMSE) of 1.945 N (vs. 2.004 N; 6D force–torque sensor-based) for end-effector force estimation and 3.006 Nm (vs. 3.043 Nm; 6D force–torque sensor-based) for robot joint torque estimation. The obtained results for an array of 1D strain gauges were comparable with those obtained with a robot’s built-in sensor, demonstrating the validity of the proposed approach.

Design, optimization, and crosstalk analysis for robot joint torque sensor of shear beam structure.

Crosstalk resistance is an important criterion for evaluating the measurement error of the Joint Torque Sensor (JTS) in actual collaborative robot application, but few research literature studies on the crosstalk resistance of shear beam-type JTS have been found. This paper proposes a mechanical structure of one shear beam sensor and determines its strain gauge working area. Multi-objective optimization equations are established with three major performance indicators of sensitivity, stiffness, and crosstalk resistance. Optimal processing and manufacturing structure parameters are obtained by employing both the response surface method based on the central composite design experimental principle and the multi-objective genetic algorithm. By simulating and experimenting, the optimized sensor is verified and has the following indices: overload resistance 300% F.S., torsional stiffness 503.44 KN m/rad, bending stiffness 142.56 KN m/rad, range 0-±200 N m, sensitivity 25.71 mV/N m, linearity 0.1999%, repeatability error 0.062%, hysteresis error 0.493%, measurement error less than 0.5% F.S. under Fx (392.4 N) or Fz (600 N) crosstalk load, and measurement error less than 1% F.S. under My (25 N m) moment crosstalk. The proposed sensor possesses good crosstalk resistance and especially axial crosstalk resistance and has good overall performance to meet well the engineering requirements.

Design and Optimization of a Joint Torque Sensor for Lightweight Robots

Force detection is an important indicator of a robot's interaction with the working environment. Therefore, accurate and sensitive torque sensor design and development are important. This study proposed a robot joint torque sensor design and optimization based on a strain gauge and through-hole spoke structure topology to detect the external force with high sensitivity and low noise. A through-hole spoke type with a novel inclination sensing surface (ISS) was proposed, and the optimization design improved the strain gauge placement position and focused stress arrangement. Three optimization objectives were desired to enhance the joint torque sensor sensitivity with guaranteed stiffness. A sensor signal processing circuit was also introduced to obtain low-noise output signal performance. The experimental results showed that the joint torque sensor had a high sensitivity of 1.65 mV/Nm before amplification. This sensor was compared with a previous study and a commercial torque sensor. The prototype torque sensor indicated high sensitivity and low error.

Control of a nonanthropomorphic exoskeleton for multi-joint assistance by contact force generation

In this article, a novel controller for a nonanthropomorphic exoskeleton robot was designed to reduce joint torque of its operator using the contact force between them. Since the joints of the nonanthropomorphic exoskeletons are not directly connected to those of the operator due to the difference between their kinematic structure, joint assistance is performed by transmitting the contact force on their coupling parts instead of transmitting the joint torque of the nonanthropomorphic exoskeleton directly into the human joint. Most of the previous studies have focused on reducing the measured contact force by moving the coupling parts or commanding the robot joint torque. On the contrary, the proposed method focuses on reducing the human joint torque, which is estimated by formulating inverse dynamics, by obtaining possible contact force solutions. The commanding torque of the nonanthropomorphic exoskeleton was calculated by inverse dynamics based on the model information. To verify the control performance of the proposed method, we have developed a simulation environment for a lower-limb nonanthropomorphic exoskeleton. When the coupling part was implemented to be rigid for an ideal case, the joint torque of the human model to perform the same motion was successfully reduced by the given torque reduction ratio. For a more realistic condition, a nonrigid coupling was also implemented as a virtual spring-damper system, and its effect on the control performance was demonstrated in the simulation.

A low cost 3D-printed robot joint torque sensor

Technological advances allow researchers to develop advanced arm robots and can safely work side by side with humans Therefore, a robot arm controller can be designed in such way that the robot arm can move along the desired trajectories and act upon external influences, in this last case, the torque sensor plays an important rule. Currently torque sensors are available in the market has a high price. In this work, an inexpensive robot joint torque sensor is presented. Most parts of this sensor are made using 3D printers. While the other components are easily can be found in the market and with a relatively low-costs. The development of this sensor is intended to facilitate the prototyping of the robot arm for educational and research purposes. The basic idea of the sensor mechanism is to convert torque into a force absorbed by a spring. Then, the encoder senses the direction and the value of the input torque. This torque sensor can be easily too customized. Thus this sensor can be tailored to the needs by replacing some parts such as encoder and spring. The mechanism of this sensor can also be adjusted with the actuator to be paired. Experiments have been conducted to verify the accuracy and the performance of the proposed torque sensor.

Adaptive torque estimation of robot joint with harmonic drive transmission

Robot joint torque estimation using input and output position measurements is a promising technique, but the result may be affected by the load variation of the joint. In this paper, a torque estimation method with adaptive robustness and optimality adjustment according to load variation is proposed for robot joint with harmonic drive transmission. Based on a harmonic drive model and a redundant adaptive robust Kalman filter (RARKF), the proposed approach can adapt torque estimation filtering optimality and robustness to the load variation by self-tuning the filtering gain and self-switching the filtering mode between optimal and robust. The redundant factor of RARKF is designed as a function of the motor current for tolerating the modeling error and load-dependent filtering mode switching. The proposed joint torque estimation method has been experimentally studied in comparison with a commercial torque sensor and two representative filtering methods. The results have demonstrated the effectiveness of the proposed torque estimation technique.

Optimal limb length ratio of quadruped robot minimising joint torque on slopes

This paper aims to determine an optimal structure for a quadruped robot, which will allow the robot's joint torque sum to be minimised. An animal's characteristic limb length ratio is a vital part of its overall morphology and the one that enables it

On controlling robots with redundancy

Recently there has been considerable interest in increasing the applicability and utility of robots by developing manipulators which possess kinematic and/or actuator redundancy. This paper presents a unified approach to controlling these redundant robots. The proposed control system consists of two subsystems: an adaptive position controller which generates the Cartesian-space control force F ϵR m required to track the desired end-effector position trajectory, and an algorithm that maps this control input to a robot joint torque vector T ϵR n . The F → T map is constructed so that the robot redundancy (kinematic and/or actuator) is utilized to improve the performance of the robot. The control scheme does not require knowledge of the complex robot dynamic model or parameter values for the robot or the payload. As a result, the controller is very general and is computationally efficient for on-line implementation. Computer simulation results are given for a kinematically redundant robot, for a robot with actuator redundancy, and for a robot which possesses both kinematic and actuator redundancy. In each case the results demonstrate that accurate end-effector trajectory tracking and effective redundancy utilization can be achieved simultaneously with the proposed scheme.