Inspiration
We wanted to create a smart, interactive companion that seamlessly integrates into daily life, promoting healthy habits while also assisting with object identification. Inspired by the growing need for proactive health monitoring and accessibility tools, we envisioned Baxter as an AI-powered assistant that encourages good posture, provides good general-health advice, and overall well-being.
What it does
Baxter is an autonomous health companion that follows users, encourages physical activity, tracks posture , and assists with object retrieval. It provides gentle reminders to stay active and stretch by following you around while using AI-powered vision to recognize and identify objects.
How we built it
We developed Baxter by combining a NAO bot with a turtle bot. We leveraged Multi modal LLM's from DeepSeek (Janus) as well as Google's gemini. The turtle bot is programmed using ROS in C++ and the NAO pot is programmed in python. The recognition software is programmed using OpenCV and Pytorch. The TTS and STT was created using PyAudio and Whisper.
Challenges we ran into
Fine-tuning Baxter’s ability to follow users smoothly without errors. Fixing the algorithm that allow for the bot to follow you Ensuring seamless real-time processing with minimal lag.
Accomplishments that we're proud of
Working together as a team and successfully within the time constraints Implementing the tracking feature Implementing the TTS and STT feature
What we learned
The importance of real-time AI processing in robotics. How to optimize tracking and navigation for dynamic environments. The potential impact of AI-driven health assistants in everyday life.
Log in or sign up for Devpost to join the conversation.