Path Planning for Robot and Pedestrian Simulations

The thesis is divided into two parts. The first part presents a new proposed method for solving the path planning problem to find an optimal collision-free path between the starting and the goal points in a static environment. Initially, the grid model of the robot's working environment is constructed. Next, each grid cell's potential value in the working environment is calculated based on the proposed potential function. This function guides the robot to move toward the desired goal location, it has the lowest value at the goal location, and the value increase as the robot moves further away. Next, a new method, called Boundary Node Method (BNM), is proposed to find the initial feasible path. In this method, the robot is simulated by a nine-node quadrilateral element, where the centroid node represents the robot's position. The robot moves in the working environment toward the goal point with eight-boundary nodes based on the boundary nodes' characteristics. In the BNM method, the initial feasible path is generated from the sequence of the waypoints that the robot has to traverse as it moves toward the goal point without colliding with obstacles. The BNM method can generate the path safely and efficiently. However, the path is not optimal in terms of the total path length. An additional method, called Path Enhancement Method (PEM), is proposed to construct an optimal or near-optimal collision-free path. The generated path obtained by BNM and PEM may contain sharp turns. Therefore, the cubic spline interpolation is used to create a continuous smooth path that connects the starting point to the goal point. The performance of the proposed method is compared with the other path planning methods in terms of path length and computational time. Moreover, the multi-goal path planning problem is investigated to find the shortest collision-free path connecting a given set of goal points in the robot working environment. Furthermore, to verify the performance of the proposed method, several experimental tests have been performed on the e-puck robot with different obstacle configurations and various positions of goal points. The experimental results showed that the proposed method could construct the shortest collision-free path and direct the real physical robot to the final destination point. At the end of the first part of the thesis, we investigate the multi-goal path planning problem for the multi-robot system such that several robots reach each goal. In the second part of this thesis, we proposed a new method for simulating pedestrian crowd movement in a virtual environment. The first part of this thesis concerning the generation of the shortest collision-free path is used. In this method, we assumed that the crowd consists of multiple groups with a different number and various types of pedestrians. In this scenario, each group's intention is different for visiting several goal points with varying sequences of the visit. The proposed method uses the multi-group microscopic model to generate a real-time trajectory for each pedestrian navigating in the pedestrianized area of the virtual environment. Additionally, an agent-based model is introduced to simulate pedestrian' behaviours. Based on the proposed method, every single pedestrian in each group can continuously adjust their attributes, such as position, velocity, etc. Moreover, pedestrians optimize their path independently toward the desired goal points while avoiding obstacles and other pedestrians in the scene. At the end of this part of the thesis, a statistical analysis is carried out to evaluate the performance of the proposed method for simulating the crowd movement in the virtual environment. The proposed method implemented for several simulation scenarios under a variety of conditions for a wide range of different parameters. The results showed that the proposed method is capable of describing pedestrian' behaviours in the virtual environment.