Skip to content
Robolist.ai
Bandit

Overview

Bandit is a friendly, socially assistive humanoid robot. It's been used to help children with autism, motivate older adults in physical exercise, and provide therapies to stroke patients. The goal of the Bandit project was developing a humanoid robot for research in socially assistive robotics, including interaction with children and adults in school, home, office, and hospital settings. Led by Maja Mataric, a team at the Interaction Lab at University of Southern California in Los Angeles and BlueSky Robotics designed a robot that is safe, expressive, and affordable. Bandit uses joint friction to hold the arms in a desired pose and a small amount of power to move to new poses. This limits its manipulation skills but lets it gesture and interact safely with people. Its relatively small size means users are not intimidated, and its face is simple, so it is easy to relate to and understand. The first prototype was created in 2004; the researchers designed an improved version in 2006 and 2007. In 2007 and 2008 they built six copies of the new design, which are the robots they use in their studies today.

Key facts

Weight
7.2 kg

Detailed specifications

Motion & kinematics2
Dof Total
15
Dof Per Arm
5
Sensors1
Sensor Suite
infrared, ultrasound, touch
Compute2
Ros Compatible
true
Onboard Compute
High-Performance Compute Unit
Other16
Height Mm
1066
Price Tier
<10K
Applications
warehouse_transport, assembly, research
Sub Category
socially assistive
Api Available
true
Datasheet Url
https://www.seeing-systems.com/wp-content/uploads/2026/02/Bandit-FPV-Specification.pdf
Pricing Model
purchase
Operating Temp
-20°c to +75°c
Company Country
US
Navigation Type
path planning
Deployment Count
2
Deployment Notes
The Interaction Lab currently owns 2 iPals.
Availability Status
available
Programming Interface
code_python
Additional Information
- Modular software platform allowing capability expansion. - Works with any end effector and integrates seamlessly with existing systems. - Core SDK enables a natural language prompt to identify objects and generate 3D pick points. - Installations can be done without extensive technical setup, eliminating the need for cloud dependencies.
Deployment Environment
indoor

Reviews for Bandit

Loading reviews…