Robot multi-task motion implementation method based on sequential logic, medium and equipment

Abstract

According to the robot multi-task motion implementation method based on the sequential logic, the medium and the equipment, the traditional TLTL is expanded to the wTLTL (Weighted Trunked Linear Temporal Logic, hereinafter referred to as wTLTL) according to the method, and the DMPs and the wTLTL are combined, so that complex operation tasks with user preferences are achieved. The DMPs are used for approximating a track with user preference, the shape of the track is determined by parameters of the DMPs, and optimization is carried out through a black box optimization strategy (hereinafter referred to as PIBB) based on a weighted average covariance matrix. The wTLTL provides an interpretable task specification, a cost function with user preference can be designed according to quantitative semantics, an improved PIBB algorithm is utilized, then the optimal shape parameter of the DMPs is recognized, and then the motion trail of a complex operation task is achieved. The user preferences are embodied by adjusting weights in the wTLTL. Compared with a traditional multi-task operation motion planning method, the method has the advantages that the motion trails of the complex operation tasks with user preferences can be quickly planned, the calculation complexity is low, and the planning efficiency is high.

Description

technical field [0001] The invention relates to the technical field of robot motion planning, in particular to a method, medium and equipment for implementing multi-task motion of a robot based on sequential logic. Background technique [0002] With the continuous development of robotics technology, robots are gradually being used in offices, hospitals, and homes. These application scenarios usually require robots to be able to autonomously learn, plan, and execute complex operational tasks consisting of a series of logical actions. [0003] Dynamic movement primitives (Dynamic movement primitives, hereinafter referred to as DMPs), as a flexible way to express motion trajectories, are widely used in the learning and planning of complex motion trajectories of robots, such as humanoid biped robots and dual-arm manipulating robots. DMPs are generally encoded as a combination of a simple linear dynamical system and a nonlinear component, and the optimal parameters of the nonline...

AI Technical Summary

Problems solved by technology

However, STL specifies subtasks with clear time boundaries, while many manipulation tasks only require subtasks to be performed in the required order, manually assigning time boundaries to subtasks may cause the robot to fail to find satisfying tasks due to unexpected environmental events. The strategy of TLTL; while the robustness of TLTL only considers the degree of satisfaction of the TLTL formula in the most extreme case, it may not be possible to find a more robust solution, and at the same time, it cannot express user preferences

Images