关键词: 养老机器人, 多任务路径规划, 引导学习, 鲸鱼算法, 强化学习, TP242（机器人）

Abstract: The global issue of aging is becoming increasingly severe, and the field of elderly care is facing a challenge of manpower shortage, urgently requiring the introduction of robot technology with intelligent decision-making capabilities. To address the autonomous path planning of assistive robots under a multi-task mechanism in elderly care scenarios, an improved Soft Actor-Critic (SAC) reinforcement learning decision algorithm was proposed. Firstly, an obstacle contour reconstruction method based on virtual circles was introduced, which reduced the complexity of environmental modeling while enhancing radar detection efficiency. Then, to tackle the difficulty reinforcement learning algorithms in optimizing strategies from scratch when solving complex tasks in a continuous state space, the Whale Optimization Algorithm (WOA) was integrated with the SAC algorithm. By constructing an auxiliary supervision mechanism to provide directional guidance for the learning process, the decision-making capability was improved while significantly accelerating the convergence speed. Finally, task planning was conducted based on the daily needs of the elderly, and model training was completed in environments that included fixed tasks with static and dynamic obstacles as well as emergent random tasks. Simulation results demonstrated that, compared to the traditional SAC algorithm, the WOA-SAC algorithm reduced the average path length from 48m to 43m, increased the success rate from 75% to 80%, and decreased the average step count from 270 to 190. These improvements highlighted the significant enhancement in the learning efficiency and decision-making capability of the SAC algorithm, effectively addressing the autonomous path planning challenges within multi-task mechanisms.

Key words: elderly care robots, multi-task path planning, guided learning, Whale optimization algorithm, reinforcement learning, TP242（robot）