A multi strategy bidirectional RRT* algorithm for efficient mobile robot path planning.

Efficient navigation in cluttered spaces remains a fundamental challenge for autonomous systems. Prior research has shown that the Rapidly-exploring Random Tree Star (RRT*) algorithm provides a robust framework for finding optimal paths in high-dimensional configurations. Standard implementations often struggle with slow convergence rates when navigating through narrow passages or dense obstacle fields. Conventional sampling methods frequently produce jagged trajectories that require significant computational overhead to refine for real-world application. Existing bidirectional approaches attempt to accelerate discovery but often lack the adaptive logic needed for truly complex spatial constraints. The lack of trajectory continuity in basic sampling-based planners often leads to poor dynamic controllability for mobile robotic platforms. This absence of evidence motivated the development of a more sophisticated, multi-strategy approach to enhance both speed and trajectory smoothness.

This study introduces the Multi Strategy Bidirectional RRT* (MS-BI-RRT*) algorithm to overcome the inherent limitations of traditional sampling-based planners. The researchers sought to integrate adaptive expansion logic that responds to the specific geometric challenges of a given environment. By implementing a dynamic scheduling mechanism, the team aimed to minimize the number of redundant nodes generated during the exploration phase. Another objective involved refining the cost function to ensure that the final path considers multiple factors beyond simple Euclidean distance. The project focused on achieving a 100% success rate across diverse environmental layouts, ranging from open spaces to intricate mazes. Developers intended to enhance the stability of tree expansion by adjusting step sizes according to the proximity of obstacles. Ultimately, the work intended to provide a solution that balances computational efficiency with the physical constraints of mobile robotic hardware.

The investigative team developed the MS-BI-RRT* framework by incorporating a bidirectional tree growth strategy that samples from both start and goal configurations. They engineered an expansion mode scheduling mechanism that utilizes dynamic goal bias probability and expansion feedback to switch between search behaviors. To maintain stability in cluttered regions, the algorithm employs a dynamic step size adjustment method calculated based on local obstacle density. During the parent node rewiring phase, the researchers utilized a multi-factor path cost function to evaluate and optimize the connectivity of the tree. The post-processing stage involves the application of a Bézier curve-based smoothing strategy to transform discrete nodes into continuous trajectories. Performance was rigorously evaluated through simulations across five distinct typical environments, comparing the new method against five established baseline algorithms. The comparative analysis specifically measured execution time, node count, and path length against BI-RRT*, APF-RRT*, BI-APF-RRT*, and GB-RRT*.

The MS-BI-RRT* algorithm achieved a remarkable 77.50% reduction in average execution time compared to the standard RRT* baseline. Data analysis revealed that the proposed method decreased the total number of nodes required for path discovery by 76.41% relative to the traditional RRT* algorithm. In terms of geometric efficiency, the new approach shortened the final path length by 4.37% compared to the baseline RRT*. The system demonstrated exceptional reliability by maintaining a 100% success rate across all five simulated environmental test cases. Comparative benchmarks against BI-RRT*, APF-RRT*, BI-APF-RRT*, and GB-RRT* consistently favored the multi-strategy bidirectional approach in every performance metric. Statistical results confirmed that the adaptive expansion mode scheduling significantly improved the convergence speed in complex obstacle configurations. These findings indicate that the integration of expansion feedback and dynamic step sizing significantly accelerates the convergence toward an optimal solution.

The proposed MS-BI-RRT* framework establishes a new benchmark for efficiency in the field of mobile robot path planning. By addressing the convergence bottlenecks of sampling-based methods, this research enables more responsive navigation for autonomous mobile robots. The inclusion of Bézier curve-based smoothing ensures that the generated paths are immediately compatible with the dynamic controllability requirements of physical platforms. Enhanced environmental adaptability suggests that this algorithm can be deployed in diverse industrial and service robotics applications without extensive tuning. The significant reduction in node count and execution time highlights the algorithm's potential for real-time implementation on resource-constrained hardware. Future developments may focus on extending these multi-strategy principles to three-dimensional environments or dynamic obstacle avoidance scenarios. The study's authors conclude that the combination of bidirectional search and adaptive expansion logic provides the robustness necessary for complex real-world navigation.