KDAIL - Autonomous Bot with Ultra-Precise UWB Localization

Project Overview

During our internship at Krishna Defence and Allied Industries, our team of 4 developed an autonomous navigation bot similar to TurtleBot with a revolutionary approach to localization. Instead of traditional encoders, we implemented Ultra-Wideband (UWB) sensors to achieve centimeter-level precision in indoor environments.

See It In Action

Testing the bot in a 10m × 10m area with precise UWB localization

The Challenge

The project aimed to solve precise localization in large areas with the following objectives:

Stage 1 (VJTI): 2cm accuracy in 10m × 10m area using UWB sensors
Stage 2 (IIT Gandhinagar): 5cm accuracy in 500m × 500m area - View Next Phase →
Replace traditional encoder-based localization with UWB technology
Develop custom navigation algorithms without relying on Nav2
Build complete firmware from scratch

Hardware Setup

Custom-built autonomous bot with UWB sensors and navigation systems

Ultra-Wideband (UWB) sensors mounted on the bot for precise localization

Revolutionary Approach - UWB Over Encoders

Unlike traditional robots that rely on wheel encoders for localization, I implemented Ultra-Wideband (UWB) sensor technology which provided several advantages:

Encoder-Free Navigation: Eliminated cumulative error from wheel slippage
2cm Precision: Achieved exceptional accuracy in 10m × 10m operational range
Real-time Positioning: Continuous absolute position without drift
Robust Performance: Works reliably in various environmental conditions

Development Process

Bot teleoperation demo showing TurtleBot-like movement capabilities

Custom Implementation

We built the entire system from ground up without relying on existing navigation stacks:

Custom SLAM: Implemented SLAM algorithms from scratch for mapping and localization
No Nav2 Dependency: Created custom navigation scripts for path planning and execution
Complete Firmware: Wrote all firmware code for the bot's control systems
UWB Integration: Developed custom drivers and algorithms for UWB sensor integration

System Testing

KDAIL navigation system in action during testing phase

Key Capabilities

Object Detection: Identifies target boxes and objects in the environment
Autonomous Navigation: Plans and executes paths to reach target locations
Precise Localization: Maintains 2cm accuracy throughout the operational area
Real-time Processing: Processes UWB signals and navigation commands in real-time
Environmental Adaptation: Works both indoors and outdoors with consistent accuracy

Built With

Ultra-Wideband (UWB) Custom SLAM ROS2 Computer Vision Custom Navigation Object Detection Firmware Development Path Planning

Technical Achievements

2cm Accuracy: We achieved exceptional localization precision in 10m × 10m area
10m UWB Range: UWB sensors provided reliable coverage across the operational area
Custom SLAM Implementation: Built mapping and localization algorithms from scratch
Encoder-Free Design: Revolutionary approach eliminating traditional encoder dependencies

Challenges Overcome

Eliminating Encoder Dependency: Successfully replaced traditional encoder-based localization with UWB technology for drift-free positioning
Custom Algorithm Development: Built SLAM and navigation systems from scratch without relying on existing Nav2 framework
Precision Requirements: Achieved demanding 2cm accuracy through careful calibration and algorithm optimization
Real-world Testing: Validated system performance in both controlled and outdoor environments

Our Robotics Journey

This project at Krishna Defence and Allied Industries marked a significant milestone in our robotics journey. Working as a team of 4, the experience of building a complete autonomous system from scratch taught us:

Advanced sensor integration and custom driver development
Real-world robotics challenges and solutions
Precision control systems and algorithm optimization
Industrial-grade system design and implementation
Working with cutting-edge UWB localization technology