Mini-Project: Robotic Nervous System

Objective

Design a sensor-driven control system that mimics the human nervous system's reflex arc using ROS 2.

---

Project Overview

Scenario:

Your humanoid robot must react to external stimuli (touch, proximity, vision) by triggering appropriate motor responses without central planning (similar to biological reflexes).

Requirements:

1. Sensor Nodes (3 minimum):

   • Touch sensor (simulated or real GPIO)
   • Proximity sensor (ultrasonic or LIDAR)
   • Camera-based object detection (OpenCV)

2. Actuator Nodes (2 minimum):

   • Servo motor controller (gripper or arm)
   • LED/buzzer feedback system

3. Reflex Logic Node:

   • Subscribes to sensor topics
   • Publishes to actuator topics
   • Implements rule-based or ML-based decision logic

---

System Architecture

graph TD
    A[Touch Sensor Node] -->|/sensor/touch| D[Reflex Controller]
    B[Proximity Sensor Node] -->|/sensor/distance| D
    C[Camera Node] -->|/sensor/vision| D
    D -->|/actuator/servo| E[Servo Controller]
    D -->|/actuator/buzzer| F[Buzzer Node]

---

Implementation Steps

Step 1: Workspace Setup

mkdir -p ~/ros2_ws/src
cd ~/ros2_ws/src
ros2 pkg create --build-type ament_python nervous_system
cd ~/ros2_ws
colcon build
source install/setup.bash

Step 2: Sensor Node Example (Proximity)

import rclpy
from rclpy.node import Node
from std_msgs.msg import Float32
import random

class ProximitySensor(Node):
    def __init__(self):
        super().__init__('proximity_sensor')
        self.publisher = self.create_publisher(Float32, '/sensor/distance', 10)
        self.timer = self.create_timer(0.5, self.publish_distance)
    
    def publish_distance(self):
        msg = Float32()
        msg.data = random.uniform(0.1, 2.0)  # Simulated distance in meters
        self.publisher.publish(msg)
        self.get_logger().info(f'Distance: {msg.data:.2f}m')

def main(args=None):
    rclpy.init(args=args)
    node = ProximitySensor()
    rclpy.spin(node)
    node.destroy_node()
    rclpy.shutdown()

Step 3: Reflex Controller Node

import rclpy
from rclpy.node import Node
from std_msgs.msg import Float32, Bool

class ReflexController(Node):
    def __init__(self):
        super().__init__('reflex_controller')
        
        # Subscribers
        self.create_subscription(Float32, '/sensor/distance', self.distance_callback, 10)
        self.create_subscription(Bool, '/sensor/touch', self.touch_callback, 10)
        
        # Publishers
        self.servo_pub = self.create_publisher(Float32, '/actuator/servo', 10)
        self.buzzer_pub = self.create_publisher(Bool, '/actuator/buzzer', 10)
    
    def distance_callback(self, msg):
        if msg.data < 0.3:  # Object too close
            self.get_logger().warn('Obstacle detected! Activating reflex...')
            self.activate_buzzer()
    
    def touch_callback(self, msg):
        if msg.data:  # Touch detected
            self.get_logger().info('Touch reflex triggered!')
            self.move_servo(90.0)
    
    def activate_buzzer(self):
        msg = Bool()
        msg.data = True
        self.buzzer_pub.publish(msg)
    
    def move_servo(self, angle):
        msg = Float32()
        msg.data = angle
        self.servo_pub.publish(msg)

def main(args=None):
    rclpy.init(args=args)
    node = ReflexController()
    rclpy.spin(node)
    node.destroy_node()
    rclpy.shutdown()

---

Deliverables

1. Source Code (50%)

• GitHub repository with:
  • All ROS 2 nodes
  • package.xml and setup.py
  • Launch file for entire system

2. Documentation (30%)

• README.md with:
  • System architecture diagram
  • Installation instructions
  • Usage examples
  • Video demo link

3. Demo Video (20%)

• Record a 3-5 minute video showing:
  • System startup
  • Sensor triggers
  • Actuator responses
  • Edge case handling

---

Evaluation Criteria

Criterion	Points	Description
Functionality	40	All sensors and actuators work correctly
Code Quality	20	Clean, commented, follows ROS 2 conventions
Architecture	20	Modular design, proper topic naming
Documentation	15	Clear README, architecture diagram
Creativity	5	Bonus for unique sensor/reflex combinations

Total: 100 points

---

Advanced Extensions (Optional)

1. Machine Learning Reflex:

   • Train a small neural network to classify sensor patterns
   • Deploy using ONNX Runtime

2. Multi-Robot Coordination:

   • Have two robots communicate reflexes via ROS 2 topics
   • Example: Robot A detects danger → warns Robot B

3. Real Hardware:

   • Deploy to Raspberry Pi + Arduino
   • Use actual GPIO sensors and servos

---

Submission Guidelines

Due: End of Week 5
Format: GitHub repository link
Include:

• Source code
• README.md
• demo_video.mp4 (upload to Google Drive/YouTube)

Submit via: Course LMS

---

FAQ

Q: Can I use C++ instead of Python?
A: Yes, but provide equivalent functionality.

Q: Is simulation acceptable or must I use real hardware?
A: Simulation is acceptable. Use Gazebo or custom mock nodes.

Q: Can I work in pairs?
A: No, this is an individual project. However, you may discuss concepts with peers.

Q: What if my sensor node crashes?
A: Implement error handling and restart logic. Document failure modes in README.