EvoRobotics Group Assignment

Team Members

• Minsol Kim
• Yu Zeyuan

Overview

This repository hosts our group's submission for Task Sheet 1 of the Evolutionary Robotics course. The submission includes all necessary code, data visualizations (plots), and supplementary resources essential for evaluating our implementation of specified robotic behaviors and sensor integrations.

Tasks and Subtasks

Completed Tasks

Task 1.1: Prepare the Robot Simulator for Braitenberg Vehicles

• Subtask 1.1: Implemented a torus space in the controller.

  • Modified both Agressor_controller.py and Fear_controller.py to support toroidal wrapping of space.

  def torus(self):
      # Retrieve current position and adjust according to the torus world rules
      robot_position_x, robot_position_y, robot_heading = self.agent.get_position()
      robot_position_x = robot_position_x % self.config['world_width']
      robot_position_y = robot_position_y % self.config['world_height']
      self.agent.set_position(robot_position_x, robot_position_y, robot_heading)

• Subtask 1.2: Integrated light sensors in my_world.

  • Established a method to define light distribution in the environment:

  def defineLight(self):
      light_source_position = np.array([self.config['world_width'] / 2, self.config['world_height'] / 2])
      max_light_intensity = 60
      light_dist = np.zeros((self.config['world_width'], self.config['world_height']))
      for x in range(self.config['world_width']):
          for y in range(self.config['world_height']):
              distance = np.linalg.norm(np.array([x, y]) - light_source_position)
              light_dist[x, y] = max(0, max_light_intensity - distance/3)
      return light_dist

Task 1.2: Implementation of Braitenberg Vehicles

• Subtask 1.1: Created light sensors for each side (left and right).
  • Developed modules light_sensor_L.py and light_sensor_R.py containing the class implementations LightIntensitySensor_L and LightIntensitySensor_R.
• Subtask 1.2: Developed two types of controllers: Aggressor and Fear.
  • For the Aggressor vehicle, the robot's behavior is guided by differential light intensity:

  sensor = self.agent.get_perception()
  delta_s = abs(sensor[1] - sensor[2])
  base_speed = 5
  sensitivity = 50
  speed = base_speed + sensitivity * delta_s
  if sensor[1] > sensor[2]:
      turn_angle = 5
  elif sensor[1] < sensor[2]:
      turn_angle = -5
  else:
      turn_angle = 0

  • The Fear vehicle utilizes the opposite strategy for turn angles.

Incomplete Tasks

• Once Agressor Vehicle got close to light source, it keeps wandering around light source. This prevent Agressor Vehicle from getting out of light source and driving the torus around freely.

Additional Documentation

• Plots: All plots are stored in the plots/ directory in PNG format. These plots illustrate light distribution and the trajectories of both types of vehicles.
• Videos: Demonstrative videos showcasing the behavior of each vehicle type are available in the videos/ directory.

Additional Notes

---