Multi-dimensional force sensor calibration decoupling method based on particle swarm optimization BP neural network

Abstract

The invention discloses a multi-dimensional force sensor correction and decoupling method based on a particle swarm optimization BP neural network. The BP network is optimized by using a global searchoptimal characteristic of PSO, the defect that the BP algorithm is trapped in local optimum in learning is avoided, and the mapping relationship of calibration data of a conversion channel is captured, so that the decoupling model has good convergence and adaptability, and the multi-dimensional force sensor has good measurement precision. By establishing the neural network model, the limitation of multi-coupling and highly nonlinear relationship of force or moment of each dimension is broken through, the decoupling capability of the multi-dimensional force sensor is improved, and the practicability of the multi-dimensional force sensor is improved; the BP network is improved and optimized by utilizing the global search optimal characteristic of the particle swarm algorithm, so that the defect that the BP algorithm is trapped in local optimum in learning is avoided, and the model has good convergence and adaptability and has good generalization ability. Compared with a traditional method, the multi-dimensional force sensor decoupling method provided by the invention has better decoupling precision and better decoupling performance.

Description

technical field [0001] The invention relates to the field of multidimensional force sensor calibration and decoupling methods, in particular to a multidimensional force sensor calibration and decoupling method based on particle swarm optimization BP neural network. Background technique [0002] With the development of the times, single-dimensional force sensors can no longer meet the measurement requirements of daily production and research, so multi-dimensional force sensors are increasingly favored by engineering applications. A multidimensional force sensor refers to a force sensor that can measure force and moment components in more than two directions at the same time. In the Cartesian coordinate system, force and moment can be decomposed into three components respectively. The most complete form of multidimensional force is six-dimensional force Or six-dimensional torque sensor, a sensor capable of simultaneously measuring three force components and three torque compon...

AI Technical Summary

Problems solved by technology

[0003] However, due to the integrated elastic body structure of the multi-dimensional force sensor and the error in the manufacturing process, there is an inevitable coupling on the conversion channels of the sensor in different dimensions, that is, inter-dimensional coupling, which seriously affects the measurement of the sensor. Accuracy, does not meet the actual measurement requirements, so it is very important to decouple the sensor

The traditional method uses a decoupling method based on the least squares method. The numerical accuracy of this method is not high, the robustness is poor, and it is easy to generate a local optimal solution. Good results have been achieved in wind power prediction, circuit system analysis, noise control, etc. The present invention proposes a multi-dimensional force sensor correction and decoupling method based on particle swarm optimization BP neural network, using PSO (particle swarm algorithm) to search for the optimal The characteristics of the BP network are optimized to avoid the disadvantages of the BP algorithm falling into local optimum during learning, and the mapping relationship of the calibration data of the conversion channel is captured, so that the decoupling model has good convergence and adaptability, and the multi-dimensional force sensor has a relatively good performance. good measurement accuracy

Images