AI-Powered Remote Control Car with voice commands
Budget: ₹12,500 – ₹37,500 INR
I want to turn a standard RC chassis into a smart, self-aware toy that can do far more than just follow joystick commands. The end goal is a prototype car that:
• Detects obstacles in real time
• Plans an optimal path around them
• Switches between full autonomous driving and classic “manual” mode on demand
Everything should be directed from a single mobile app that works seamlessly on both iOS and Android. From that app I need to:
1. Watch a live video feed (low-latency)
2. Toggle between autonomous and manual modes
3. See a simple map of the car’s planned path and any detected hazards
4. Push OTA firmware updates when new AI models are trained
Hardware is already on hand (Raspberry Pi 4, camera module, ultrasonic sensors, ESC, and an ESP-based controller), so focus your bid on the software stack and integration. I’m comfortable if you choose TensorFlow Lite, PyTorch Mobile, or another lightweight framework—just explain why it’s the right fit.
Acceptance criteria
• Car navigates a 5 × 5 m obstacle course without human input, logging a collision-free run at least 3 out of 3 times.
• Manual mode latency from phone command to car response <150 ms on local Wi-Fi.
• App is published to TestFlight and Google Play Internal Testing with clear build instructions.
• All code, trained models, and wiring diagrams are delivered in a Git repo with a concise README.
If parts of the solution are modular (e.g., separate microservice for path planning), note that—I’m open to future upgrades. Let me know your approach, timeline, and any special sensors or libraries you’d recommend.
• Detects obstacles in real time
• Plans an optimal path around them
• Switches between full autonomous driving and classic “manual” mode on demand
Everything should be directed from a single mobile app that works seamlessly on both iOS and Android. From that app I need to:
1. Watch a live video feed (low-latency)
2. Toggle between autonomous and manual modes
3. See a simple map of the car’s planned path and any detected hazards
4. Push OTA firmware updates when new AI models are trained
Hardware is already on hand (Raspberry Pi 4, camera module, ultrasonic sensors, ESC, and an ESP-based controller), so focus your bid on the software stack and integration. I’m comfortable if you choose TensorFlow Lite, PyTorch Mobile, or another lightweight framework—just explain why it’s the right fit.
Acceptance criteria
• Car navigates a 5 × 5 m obstacle course without human input, logging a collision-free run at least 3 out of 3 times.
• Manual mode latency from phone command to car response <150 ms on local Wi-Fi.
• App is published to TestFlight and Google Play Internal Testing with clear build instructions.
• All code, trained models, and wiring diagrams are delivered in a Git repo with a concise README.
If parts of the solution are modular (e.g., separate microservice for path planning), note that—I’m open to future upgrades. Let me know your approach, timeline, and any special sensors or libraries you’d recommend.