This project implements the Rapidly-exploring Random Tree (RRT) algorithm for different types of robots in a 2D plane. The implementation focuses on point and line robots and explores various sampling strategies and step sizes.
- Objective: Implement RRT for an omnidirectional point robot using uniform sampling.
- Results:
- Figures 1 and 2 show results for 'shot.png' and 'simple.png' with a point robot.
- The RRT expands across the map without specific knowledge of the target position.
- Figure 3 demonstrates guiding point visualization.
- Objective: Implement RRT with Gaussian distribution centered on the target.
- Results:
- Figure 4 shows the RRT implementation for a point robot with Gaussian distribution.
- The implementation focuses more on the target area, as seen in Figure 5.
- Objective: Analyze the effect of varying step size on RRT with Gaussian distribution.
- Results:
- Figures 6 and 7 depict the number of nodes vs. step size and path length vs. step size, respectively.
- Larger step sizes lead to faster exploration but less smooth paths.
- Objective: Implement RRT for a line robot using uniform sampling.
- Results:
- Figures 8 and 9 show results for 'shot.png' and 'simple.png' with a line robot.
- The line robot avoids incorrect paths more effectively compared to the point robot.
- Objective: Study the impact of different robot lengths on the line robot's performance.
- Results:
- Figure 10 demonstrates iterations vs. robot size.
- Larger robots navigate more efficiently, as indicated by a general decreasing trend in iterations.
This project demonstrates the versatility of the RRT algorithm in robotic path planning and highlights the effects of different sampling strategies and robot configurations on its performance.