Mechanical arm tactile external force estimation method based on mechanism data hybrid model

Abstract

The invention relates to a mechanical arm tactile external force estimation method based on a mechanism data hybrid model, which comprises the following steps of: in a non-contact external force state, establishing a mechanical arm Euler-Lagrange rigid body dynamic model, and taking a difference between an actual joint torque and a mechanism model calculation torque as an unmodeled residual dynamic state; measuring and establishing a data set containing a joint state and an unmodeled residual dynamic state; training an unmodeled residual dynamic model compensation item based on the data set by using a supervised statistical learning method of Gaussian process regression to obtain a mechanism data hybrid model containing a mechanism model and the compensation item; joint external force is modeled and serves as pseudo observed quantity for dimension expansion, and a mechanical arm external force estimation model is established; and designing an external force estimation method with process noise variance updated along with the joint state. According to the method, mechanical arm tactile external force estimation with unknown coupling and unmodeled dynamics is achieved, and the method is suitable for a multi-scene multi-type force interactive mechanical arm system needing high-precision external force estimation.

Description

technical field [0001] The invention belongs to the field of electrical data processing, and in particular relates to a tactile external force estimation method of a mechanical arm based on a mechanism data mixed model. Background technique [0002] With the maturity and development of robot technology, robots have begun to replace humans in various special and even extremely harsh operating environments. In the future war rescue, space manipulators and extreme environment capture scenarios, the robot's interactive perception of the environment is related to the survival of the robot. In these unknown environments and complex tasks, real-time perception and full utilization of high-precision interaction force information is an urgent need for robots to interact intelligently and survive safely. The existing means of interactive force information perception are mainly realized by external force sensors. However, external force sensors are limited by complex structures, high c...

AI Technical Summary

Problems solved by technology

However, the above three methods are based on the simplified manipulator model, ignoring many unmodeled nonlinear dynamics coupled with external forces in the original system, and it is difficult to guarantee the estimation accuracy

In the Chinese invention patent CN201410112717.5, a method for estimating the external force of the manipulator based on H∞ filtering is proposed. Aiming at the problem of estimating the external force of the actual manipulator system, an analog low-pass filter is used to analyze the signal in the dynamic equation Perform filtering, expand the external force and estimate the external force based on H∞ filtering, without the measurement of the acceleration signal. However, this method does not consider the actual model uncertainty, which affects the accuracy of external force estimation

Chinese invention patent CN201710834869.X proposes an external force estimation method and device, which uses a multi-layer perceptron to model and compensate modeling errors, and realizes external force estimation without external force sensors. By introducing a Kalman filter, the observation In the presence of noise and model errors, it has certain robustness, but it does not consider the confidence of error modeling, and the modeling method using multi-layer perceptron is not interpretable, and the amount of data required for modeling parametric models is very large. Large, not suitable for practical engineering applications

[0004] To sum up, when the force-interactive manipulator system has unknown and unmodeled dynamics coupled with external forces, the existing methods lack a high-precision tactile force separation estimation method that can resist coupling interference, and it is urgent to overcome the data-augmented hybrid model-based infinite Sensory tactile force estimation method

Images