PIRA - Personal Indoor Robot Assistant

Date: December 10, 2024

Overview

This project centers on building a personal indoor robot assistant capable of autonomous navigation and real-time obstacle avoidance in an indoor environment. The system supports both fully autonomous waypoint traversal, guided by an indoor positioning system (Optitrack), and manual driving via WASD inputs. A Node.js server coordinates data exchange among multiple robots on the same Wi-Fi network, while a Streamlit interface provides a live 2D visualization of each robot’s position and waypoints. TingoDB stores positional data for real-time monitoring and historical analysis.

---

Key Features

• Autonomous Navigation

  • Uses feedback control (PID) to follow predefined waypoints within a ±10cm margin.
  • Leverages Optitrack for accurate (x, z, θ) positioning updates.

• Obstacle Avoidance

  • Sharp-IR sensors detect nearby obstacles.
  • Priority-based control ensures the robot diverts from its path to avoid collisions.

• Multimodal Control

  • Automatic Mode: Robot autonomously moves between waypoints, using PID corrections and sensor feedback.
  • Manual (WASD) Mode: Users can override autonomy and steer the robot via keyboard inputs sent through the Node.js server.

• Real-time Monitoring & Visualization

  • Live 2D graphics in a Streamlit dashboard show robot positions, waypoints, and obstacle triggers.
  • TingoDB logs each position update, enabling real-time data tracking and simple analytics.

• Scalable Networking

  • Node.js server handles multiple robots simultaneously.
  • Each robot maintains continuous Wi-Fi connectivity and exchanges UDP messages for minimal overhead.

---

System Requirements and Architecture

Hardware Integration

1. ESP32 Microcontroller

   • Central controller for motor actuation, sensor readings, and Wi-Fi communication.
   • Configured with FreeRTOS tasks to manage concurrency.

2. Sharp-IR Sensor

   • Analog distance sensor (ADC-based) providing obstacle detection.
   • Calibrated to detect objects within a configurable distance threshold.

3. Optitrack System

   • Provides precise localization of the robot in indoor environments.
   • Streamed position/heading data is used for navigation and control.

4. Motors & Driver

   • Controlled via PWM (LEDC on the ESP32) for linear and rotational motion.
   • Ensures fine-tuned speed adjustments, crucial for PID feedback loops.

5. Wi-Fi Networking

   • All robots connect to the same local network.
   • UDP sockets facilitate efficient real-time messaging between robots and the server.

---

Solution Design

Core Application Initialization (app_main)

• Wi-Fi Setup:
  Configures SSID, password, and reconnect parameters. Ensures each robot remains online to receive commands and transmit location updates.

• Motor Control Configuration:
  Initializes PWM channels (LEDC) for forward, backward, turning, and stopping states.

• Sharp IR Sensor & ADC:
  Sets up an ADC channel to read distance values. Calibrates the voltage-to-distance curve for obstacle detection.

• Timers and Tasks Setup:
  Creates periodic timers for tasks like checking sensor readings, computing PID corrections, and sending position requests.

Key FreeRTOS Tasks

1. Network Listener Task

   • Listens for UDP messages (e.g., WASD commands, mode toggles, new waypoint definitions).
   • Parses incoming data and updates the robot’s operating state (manual vs. autonomous).
   • In autonomous mode, processes next waypoint instructions and speed/direction updates.

2. Request Location Task (request_location)

   • Periodically sends requests to the Node.js server to obtain the robot’s latest position from Optitrack.
   • Updates internal state with (x, z, θ) data for use in PID navigation.
   • In autonomous mode, computes the heading/distance to the target waypoint.

3. Motor Control Task

   • Implements a finite state machine (FSM) for movement: STATE_MOVE_FORWARD, STATE_GO_LEFT, STATE_GO_RIGHT, STATE_STOP, etc.
   • Adjusts PWM duty cycle based on commands from manual input or PID output.
   • Manages turning maneuvers and speed scaling for smooth directional changes.

4. Obstacle Detection Task

   • Continuously reads Sharp-IR sensor values.
   • Triggers immediate avoidance behavior if an obstacle is within the critical threshold.
   • Temporarily overrides PID path corrections, prioritizing collision prevention.

5. Autonomous Navigation & PID

   • Uses proportional, integral, and derivative terms to stabilize heading toward each waypoint.
   • Computes heading error (difference between current angle and target angle).
   • Dynamically adjusts motor outputs to minimize error while respecting obstacle avoidance triggers.

---

Network Communication

Node.js Server

• Command Dispatch:
  Receives manual drive commands (WASD) or mode-change signals from a user interface.
  Broadcasts these instructions to the specified robot via UDP.

• Location Management & TingoDB:
  Aggregates position reports from each robot.
  Stores each robot’s (x, z, θ) data in TingoDB for real-time updates and historical lookback.

• Streamlit Dashboard:
  Polls the database, displaying a 2D map of robots, waypoints, and obstacles.
  Allows users to observe multiple robots’ positions and statuses simultaneously.

UDP Protocol

• Minimal overhead ensures fast, near-real-time communication for multi-robot scenarios.
• Periodic acknowledgment or keep-alive checks maintain consistent connectivity.

---

Results

• Consistent Waypoint Navigation:
  Achieved ±10cm accuracy in autonomous mode, leveraging PID corrections and accurate Optitrack data.

• Effective Obstacle Avoidance:
  Sharp-IR sensor integration prevented collisions, with priority-based logic overriding normal path following as needed.

• User-Friendly Remote Control:
  Seamless toggling between manual (WASD) commands and full autonomy.
  Real-time feedback through the Node.js server and the Streamlit interface.

• Scalable Multi-Robot Setup:
  Multiple robot assistants operate in parallel, each with independent control loops but centralized data coordination.

---

Challenges

• Real-Time Communication & Coordination:
  Managing concurrent UDP messages for multiple robots without packet collisions or data delays.

• PID Tuning:
  Ensuring smooth path tracking without oscillations required iterative parameter tuning under varying speeds and load conditions.

• Sensor Calibration:
  The Sharp-IR sensor’s accuracy can fluctuate based on environment and lighting, necessitating consistent recalibration.

---

Potential Improvements

• Interrupt-Driven Sensor Reads:
  Using hardware interrupts instead of polling for sharper real-time responsiveness.

• Path Planning Algorithms:
  Incorporate more advanced planning (e.g., A*, RRT*) for complex obstacle layouts.

• 3D Visualization or SLAM:
  Extend the system to build a richer map of the environment for real-time 3D tracking.

• Security Enhancements:
  Implement encryption for UDP messages or advanced authentication methods for sensitive deployments.

---

Demonstrations

• Functionality Demo Video
• Design Walkthrough Video

Appendix

Circuit diagram:

This personal indoor robot assistant project brings together embedded systems design, real-time networking, and automated control to achieve robust waypoint navigation and obstacle avoidance. Its flexible architecture, modular codebase, and user-friendly interface make it a compelling foundation for further robotic exploration and innovation.