Abstract
Autonomous ground vehicles operating in resource-constrained and adversarial environments face the dual challenge of maintaining operational effectiveness while minimizing detectability through intelligent navigation strategies. Current trajectory planning approaches inadequately address temporal degradation of environmental intelligence and energy optimization as fundamental design considerations. This research presents an integrated trajectory planning framework unifying energy-aware control with dynamic threat assessment through supervised autonomy principles. The framework combines two algorithmic contributions within a behavior tree architecture enabling adaptive decision-making with human oversight capabilities. The first contribution introduces AISEA-MPPI (Adaptive Importance Sampling Energy-Aware Model Predictive Path Integral), integrating terramechanics-based energy modeling with adaptive importance sampling for multi-axle autonomous vehicles. This approach demonstrates energy efficiency improvements of 6.1-10.6% across diverse terrain conditions while maintaining superior path-following accuracy and reduced actuator wear. The second contribution advances threat-aware navigation through the Enhanced Dynamic Belief (EDB) framework, operating in belief space to model temporal uncertainty decay, adversary adaptation patterns, and information staleness effects. EDB maintains 95% of optimal performance under severe information degradation where static approaches degrade to 6% effectiveness, representing a 15.8-fold improvement. Integration through behavior tree supervision enables dynamic strategy selection and human operator intervention, implementing supervised autonomy principles that maintain operational safety while maximizing autonomous effectiveness. The unified framework functions as a reactive planner receiving high-level mission directives and generating optimized trajectories suitable for real-time deployment. Physics-based simulation validation confirms applicability across mining operations, agricultural automation, disaster response missions, and security-critical scenarios requiring concealed navigation capabilities.