SmartReach

SmartReach Arm
SmartReach Arm

Inspiration

Our project, SmartReach, was born out of a desire to create meaningful technological solutions that can directly impact lives—particularly for the elderly and people with disabilities, including those living with Parkinson’s disease. These individuals often face difficulty in performing day-to-day tasks, and we wanted to create a robotic companion that could assist them in a practical, intelligent, and empathetic way.

What it does

SmartReach is an assistive robotic system powered by the SO100 robotic arm, integrated with Google’s Gemini AI model, and designed to act as a smart, helpful companion. The system can identify, locate, and retrieve objects on voice command, easing everyday tasks for individuals with limited mobility.

Here’s how it works:

The SO100 robotic arm is equipped with a Logitech webcam, and all its parts are custom 3D-printed.

Using inverse kinematics, the robotic arm calculates the precise joint angles required to reach a given object.

When a user gives a command such as, “Can you find the keys on the table?”, the system captures an image using the camera.

This image is sent to Gemini AI, which analyzes the photo to identify if the object (in this case, keys) is present.

If the object is found and the user follows up with a command like, “Can you give me the keys?”, the robot initiates the grabbing sequence, brings the object to the user, and awaits further interaction.

The idea is to have a fully responsive robotic assistant that can intelligently interpret requests and act in the physical world bridging the gap between AI and assistive robotics.

How we built it

All mechanical parts for the SO100 robotic arm were 3D-printed, allowing us to fully customize the design.

We used STS3215 servos with a single-bus connection, which simplified wiring and reduced the overall hardware complexity.

The system was programmed in Python, leveraging ROS 2 (Robot Operating System) to control the robotic arm and integrate AI-based decision-making.

Gemini AI was connected to analyze captured images and interpret user commands, acting as the "brain" of the system.

We developed the inverse kinematics logic to calculate and execute real-time movements for object retrieval tasks.

Challenges we ran into

3D printing proved to be a bottleneck—we had to frequently go back and forth to fix quality issues and reprint defective parts.

Our hardware was limited and not high-end, but we found creative solutions to make the most of what we had.

Integrating Gemini with real-time robotic controls posed a steep learning curve, but we managed to build a stable communication pipeline between the robot and the AI model.

Accomplishments that we're proud of

We always aspired to create something for social good, and we are incredibly proud of the progress we made within such a short time.

Despite limited hardware and access to professional tools, we built a fully functional prototype from scratch that surpassed our expectations.

Our team took on the challenge of learning completely new technologies, including ROS 2, inverse kinematics, and AI integration—all within the span of a hackathon.

We stepped out of our comfort zones and proved that with passion, collaboration, and determination, it's possible to turn an ambitious vision into reality.

What we learned

This was our first hands-on experience with building and controlling a robotic arm, and we gained invaluable insights into its mechanics and control systems.

We learned how to use ROS 2, a powerful framework for robotic applications.

We implemented inverse kinematics from scratch to control precise arm movements.

Most importantly, we learned how to bridge the gap between AI and robotics, turning abstract commands into tangible physical actions.

What's next for SmartReach

We plan to improve the object recognition accuracy and reduce latency in AI processing. We aim to add voice interaction to create a completely hands-free experience. Long-term, we envision SmartReach becoming a commercially viable assistive robot that can live in people’s homes and help with daily tasks—enhancing independent living for the elderly and people with physical disabilities. This project marks a small but significant step in building human-centric robotics, and we are excited to take it even further.