Intelligent Simulation System of Human-Computer Interaction Motion Based on Bullet Physics Engine

With the rapid advancement in computer technology, human motion simulation has gained substantial attention across various domains; however, existing simulation models frequently encounter challenges such as poor controllability, inadequate stability, and limited anti-interference capability.To address these limitations, this study introduces an intelligent human-computer interaction (HCI) motion simulation system based on the Bullet physics engine and reinforcement learning.The proposed system leverages the proximal policy optimization (PPO) algorithm, enhanced with a multi-objective reward function designed to precisely constrain policy updates through the comparative analysis of new and old strategies.Experimental evaluations reveal significant performance improvements, with the PPO algorithm achieving maximum reward values notably higher (by 43.5 and 307.2) than comparative algorithms, and demonstrating a minimum loss error reduction of up to 0.059.The model's stability and anti-interference capabilities were further validated under challenging scenarios involving external forces and obstacles, showing quicker recovery and superior resilience.Additionally, the model demonstrated a high correlation with real human joint movements, with fitting accuracies of 91.4% for knee joints and 93.2% for hip joints, highlighting its effectiveness and practical applicability.As a result, the study effectively improves the stability and anti-interference of the human simulation model and realizes the accurate control of the model.