Key points
- PhoneBot is an open source humanoid robot built by UCLA researchers using inexpensive Android devices for core computing and sensing tasks.
- The hardware costs approximately $400, excluding the smartphones, which are repurposed to handle vision, balance and motor control via Wi-Fi.
- Tests used four Android phones, including a 2017 Honor 9 and three budget models, to run TensorFlow Lite locomotion policies and person detection.
The short version
- UCLA researchers developed PhoneBot, a compact humanoid robot that integrates old Android smartphones as its primary computing and sensing unit.
- The system uses the phone’s camera for vision and its inertial sensors for balance, communicating with motor controllers wirelessly over Wi-Fi.
- Hardware costs for the robot body are roughly $400, making it significantly cheaper than many existing small humanoid research platforms.
The details
Researchers at the University of California, Los Angeles, created PhoneBot to demonstrate how discarded mobile devices can power robotic systems. According to TechRadar, the platform is open source and stands 483 millimeters tall while weighing 1.8 kilograms. It uses 13 servo motors to manage leg movement and torso rotation. Instead of buying dedicated cameras or expensive onboard computers, the team reused smartphones to handle vision, computing and walking control.
The robot’s architecture splits tasks between the phone and an internal computer. The smartphone captures video for person detection and pose estimation using its camera. Its built-in inertial measurement unit provides motion data that helps the robot maintain balance during learned walking movements. A separate onboard computer receives wireless commands from the phone via Wi-Fi and translates them into joint movements. This setup allows the camera to swivel toward people independently of the robot’s body orientation.
The development team tested the system with four different Android phones. These included an Honor 9 model released in 2017 and three inexpensive handsets priced between $50 and $80. All devices successfully supported core functions like motion sensing, visual detection and human interaction. The application runs locomotion policies through TensorFlow Lite. Training used reinforcement learning in simulation to reward stable movements and penalize excessive motor loading or unstable posture.
The context
This project highlights a shift toward repurposing consumer electronics for specialized hardware tasks. By leveraging the existing sensors and processing power in smartphones, researchers reduced the need for proprietary robotic components. The open source nature of the platform allows other makers to inspect and modify the design. The cost structure, with hardware excluding the phone at approximately $400, positions it as an affordable option for academic and hobbyist projects compared to traditional small humanoid systems.
The reliance on Wi-Fi for communication between the phone and the motor controller introduces a wireless dependency in the control loop. However, the source material does not detail latency issues or security implications of this wireless link. The focus remains on functional performance, specifically how the phone’s hardware enables vision-based interactions and balanced locomotion without expensive dedicated sensors.
What to watch
- Monitor updates to the open source repository for compatibility changes with newer Android versions or hardware revisions.
- Observe how the wireless communication protocol between the phone and motor controller handles interference in crowded RF environments.
- Track whether the reinforcement learning models require continuous retraining when deployed on different phone models with varying sensor specifications.
- Watch for community contributions that might expand the robot’s capabilities beyond basic walking and person tracking.
What it means for organisations
Operations teams should assess how repurposed consumer electronics impact maintenance workflows and support contracts. When hardware relies on off-the-shelf devices with varying lifecycles, replacement strategies become complex. Engineering leaders must verify that wireless control links meet reliability standards for their specific use cases. Procurement groups need to clarify who owns the software stack and whether open source licenses align with corporate policy. IT security staff should evaluate the attack surface introduced by connecting mobile devices to critical motion controllers.
What to do and how to stay safe: PhoneBot
- Audit wireless connections between control units and actuators for authentication and encryption to prevent unauthorized command injection.
- Verify that the operating system on repurposed mobile devices is updated to patch known vulnerabilities in the Android framework.
- Establish a clear inventory and lifecycle management process for consumer-grade hardware integrated into operational equipment.
- Test the system’s stability in environments with high radio frequency interference to ensure control signals remain reliable.
Step-by-step guide: Secure Home Routers for Small Business Networks
General security guidance from the TechSec Headlines newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
How much does the PhoneBot hardware cost?
The robotic hardware costs approximately $400, excluding the smartphones used for computing and sensing.
Which smartphones were used in the PhoneBot tests?
Researchers tested four Android phones, including a 2017 Honor 9 and three inexpensive models costing $50 to $80.
Does PhoneBot use dedicated cameras and computers?
No, it relies on the smartphone’s existing camera, inertial sensors and processing power for vision, balance and control.
