6249

Most robotics books either stop at the component level or jump straight to high-level frameworks that abstract the hardware away. This book occupies the middle ground: a rigorous, system-level account of how a complete autonomous platform is designed, integrated, and debugged from first principles — by a single engineer, at a real workbench, with real constraints. Every chapter reflects a problem that had to be solved and a failure that had to be understood before progress was possible.

  • Three brains, one robot. A distributed cluster of three dedicated dual-core MCUs separates perception, motor control, and navigation — eliminating the timing conflicts that cripple single-chip designs.
  • 360° awareness at 20 Hz. Sixteen time-of-flight sensors scan in parallel across 8 independent I²C buses, delivering a complete environmental sweep far faster than any sequential approach — feeding a neural network that drives real-time obstacle avoidance without human intervention.
  • AI that runs on the microcontroller itself. No cloud, no GPU — the obstacle-avoidance system runs as a trained neural network on bare metal, via TFLite Micro, in real time.
  • Fully remote-controlled from your smartphone. A built-in Wi-Fi access point and browser-based interface put real-time telemetry, sensor visualisation, and manual drive control in the palm of your hand — no app installation required.
  • Scriptable autonomous behaviour. A small onboard Virtual Machine executes bytecode scripts one instruction per scheduler tick, so long, multi-step manoeuvres run in the background without ever blocking the real-time control loops — the non-blocking foundation this book builds on toward full autonomy.
  • Toward full autonomy. The closing chapter names exactly what separates shared control from true Level 3 autonomy — self-localisation, autonomous mission planning, drift-free heading, predictive power management, and knowing when to stop — and sketches how to close each gap, before handing you the brief to design your own Black Shadow 2.0.