Back to Home

Biped Copilot

Assistive navigation for blind and visually impaired users

biped Work Project
Biped NOA devices

Overview

Biped Copilot is a wearable robotics platform that enables blind and visually impaired users to navigate independently. Using stereo depth cameras worn on the chest, it provides real-time obstacle detection, ground analysis, and spatial audio feedback through bone conduction headphones.

Over 3.5 years as a core engineer, I contributed across the full stack: from low-level camera drivers and point cloud processing to AI-powered scene descriptions and multi-language support, shipping 45+ releases across 3 hardware platforms to users in 12+ languages.

45+ Releases shipped
3 Hardware platforms
12+ Languages supported
3.5y Duration

Development Timeline

2021 Foundation
  • Multiprocessing architecture with shared memory and inter-process communication
  • Obstacle detection with YOLOv5, ground plane detection, SORT tracker
  • 3D spatialized audio feedback with priority management
  • BLE communication with companion mobile app
Nov 2022 - Mar 2023 Pre-production
  • Debian packaging and production service deployment
  • Rerun 3D visualizer integration for real-time debugging
  • Remote diagnostics upload, camera calibration and assignment scripts
  • Multi-language audio assets (French, German)
Mar - Jun 2023 Edge 2 + Navigation
  • Edge 2 hardware support with obstacle tracking and risk model
  • Image polling/processing pipeline split for performance
  • Hemispatial neglect priority, navigation commands, guide dog rule
  • Point cloud stabilization and 3D world-to-camera projection
Jun - Nov 2023 3D Perception
  • 3D Kalman tracking with speed estimation
  • Hole detection using KD-tree ground analysis
  • Small obstacle detection with Cython optimization
  • VIM3 hardware platform support, copilot as installable package
Dec 2023 - Feb 2024 GPS + Risk Model
  • GPS integration for outdoor positioning
  • Risk model with multi-Gaussian collision probability
  • Hand and occlusion detection, motion-adaptive detection range
  • Elliptical corridor rework for obstacle proximity
Mar - Apr 2024 AI Scene Description
  • OpenAI-powered scene descriptions triggered by button press
  • Text-to-speech generation with audio streaming
  • App state machine, prompts managed server-side
  • Developer tooling: monorepo merge, benchmarking integration
May - Nov 2024 NOA Device
  • NOA hardware platform: button strip, menu system, AI detection classes
  • Sentry error tracking, JWT authentication, Docker builds
  • End-to-end testing framework, Webots simulator integration
  • Obstacle audio sharpness, Gemini-generated assets
Dec 2024 - Jan 2025 Multi-language
  • 12+ language support with online TTS and auto-translate
  • Local occupancy grid for spatial mapping
  • BLE and audio system rework
  • Token-to-audio streaming, JPEG compression for AI requests
Feb - Jun 2025 Video Description + Navigation
  • Real-time video description with route following
  • Structured AI output with file caching
  • Rich navigation instructions, favorite destinations, GPX recording
  • Person finding, obstacle scanning, concurrent scene/video descriptions

Technical Scope

The system runs a Python multiprocessing architecture on embedded ARM boards (Khadas Edge 2, VIM3), managing concurrent tasks for image acquisition, obstacle detection, feedback generation, BLE communication, localization, and AI descriptions.