← index

hardware/ · 10

Smart E-Scooter (Hardware Rig + Gemini AI + Maps)

A fully functional physical e-scooter bench rig integrating Arduino controls, ProtoPie Connect, live Google Maps turn-by-turn routing, and Gemini AI voice assistance for early-stage hardware-software validation.

Fig. 1. Full e-scooter hardware bench rig with physical throttle, Arduino serial bridge, ProtoPie cluster, Google Maps API, and Gemini AI voice assistance.
Fig. 1. Full e-scooter hardware bench rig with physical throttle, Arduino serial bridge, ProtoPie cluster, Google Maps API, and Gemini AI voice assistance.

01 problem

Validating digital cluster UX for micromobility vehicles (like e-scooters and e-bikes) typically stalls until physical production hardware is assembled. Without an integrated hardware-in-the-loop testbed, design and software teams cannot evaluate how physical inputs (variable throttles, mode switches, external relays) feel alongside live UI, map navigation, and conversational AI.

02 approach

Fabricated a bench rig with physical handlebar controls (throttle potentiometer, voice-trigger push button, riding mode switch, and an external 12V LED headlight relay) driven by an Arduino.

Engineered a custom Node.js / Socket.IO bridge application to handle serial communication from the Arduino, stream live Google Maps & Places API routes, and broker WebSocket messages to ProtoPie Connect.

Integrated Gemini AI via multimodal voice APIs to process driver requests, generate real-time navigational context, and trigger proactive rerouting dialogues when low-battery thresholds are met.

Built the interactive e-scooter cluster UI in ProtoPie Studio to display live speedometer readings, mode toggles (Eco 45 km/h cap vs. Normal 90 km/h cap), and bi-directional headlight switch states.

03 stack

prototype ProtoPie Studio · E-Scooter Cluster HMI
host ProtoPie Connect · WebSocket / Socket.IO
hardware Arduino · handlebar throttle, buttons, LED relay
ai / voice Gemini AI API · voice interaction & context suggestions
maps Google Maps & Places JavaScript API
bridge Node.js bridge server · serial-to-socket relay

04 outcome

Physical throttle, mode switches, and live LED headlights bi-directionally synced with the ProtoPie digital cluster via Arduino serial
Gemini AI processes voice queries, auto-reroutes via Google Places/Directions API, and proactively suggests charging stops on simulated low battery

The rig showed that full hardware-software integration can be tested early in the development cycle. Designers and developers could evaluate the complete riding UX (physical throttle feedback, mode governance, AI voice rerouting, and headlight control) long before committing to physical manufacturing.

05 artifacts

github github.com/jkuts/eScooter_Rig
video youtu.be/FCu3iMXfHic