HGHuaquan Gan

Huaquan Gan

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Huaquan Gan has a total of 8 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Mobile robot path planning based on multi-branch network structure and evolutionary strategy of priority experience sampling TD3 algorithm

    by Lieping Zhang, Zhongtao Leng, Ming Zhang, Lanqing Shao, Huaquan Gan

    To address the issues of insufficient multi-source perception utilization and inefficient policy exploration of the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm in mobile robot path planning, an improved TD3 algorithm integrating a multi-branch perception network, evolutionary strategy, and safety-goal-oriented prioritized sampling is proposed: first, a multi-branch network architecture is designed to independently encode LiDAR data, depth camera bird's-eye view data, and robot motion states, followed by feature fusion to achieve perception decoupling and high-precision environmental representation; second, a diverse Actor policy population is constructed, where elite policies are screened based on cumulative rewards, and new individuals are generated by combining Critic-based gradient fine-tuning and adaptive parameter mutation to enhance global exploration capability; finally, a safety-goal-oriented hybrid prioritized experience replay mechanism is developed, which dynamically calculates sample priorities by integrating obstacle distance, goal proximity, and TD error to improve sample utilization efficiency. Comparative experiments in both simulation and real-world environments demonstrate that the proposed algorithm significantly outperforms comparative methods including vanilla TD3 and TD3 variants with noise-driven and multi-step update designs in terms of path length, planning time, and training stability, and exhibits excellent robustness and generalization capability in complex obstacle environments.

  • Path planning algorithm for mobile robots using value distribution and LSTM-based Soft Actor-Critic

    by Huaquan Gan, Lieping Zhang, Lanqing Shao, Shuilan Wang, Hongyuan Chen

    To address unstable value estimation and insufficient use of temporal information by the soft Actor–Critic (SAC) algorithm in dynamic environments, we propose an improved SAC-based path-planning method that integrates distributional value learning and long short-term memory (LSTM). First, the critic models the return distribution via quantile regression to reduce the variance of value estimates and improve the stability of policy evaluation. Second, an LSTM encodes historical obstacle observations to enhance prediction of environmental dynamics and decision-making. Through comparative experiments in dynamic scenarios on the Gazebo simulation platform, the proposed method outperforms LSTM-enhanced DDPG, vanilla SAC, distributional SAC, and LSTM-SAC in average path length, average steps, and average planning time. Moreover, real-world experiments further verify the feasibility and robustness of the approach.