Drone Obstacle Avoidance System

Project Overview

Developing a comprehensive obstacle avoidance system for Buckeye Vertical's competition drone, enabling autonomous navigation in dynamic environments with multiple UAVs. The system uses advanced sensor fusion and real-time path planning to ensure safe operation during high-stakes drone competitions.

360° Detection

Omni-directional obstacle detection using LIDAR technology

20 m/s

Capable of avoiding obstacles at high relative velocities

1m Accuracy

Precise object localization in 3D space

100ms Latency

Real-time processing and response

Technical Implementation

Detection System

Implemented a LIDAR-based detection system with:

40m detection range for 1m³ objects
360° horizontal field of view
Custom point cloud clustering algorithms
Vibration-resistant mounting system

Avoidance Algorithm

Developed a sophisticated avoidance system featuring:

Dynamic path planning with real-time trajectory updates
Three-dimensional velocity obstacle method
Optimization for minimal deviation from planned path
Fail-safe mechanisms and manual override capability

System Integration

Seamless integration with drone systems through:

Custom sensor data pipeline
Real-time processing on Nvidia Jetson
PX4 flight controller integration
Comprehensive testing framework

Technologies Used

LIDAR Python C++ ROS PX4 Gazebo MATLAB

Key Achievements

• Developed comprehensive simulation environment
• Achieved sub-100ms system latency
• Created detailed documentation for future teams

Project Timeline

Oct 2024

Sensor Selection & Testing

Nov 2024

Algorithm Development

Jan 2025

System Integration

Apr 2025

Final Testing & Documentation