The startup develops an in-house self-driving car platform specifically designed for low-speed environments such as university campuses and industrial parks. This technology enhances transportation efficiency by minimizing emissions and lowering accident rates in areas with limited traffic.
Funding
Funding not disclosed
Founders
Product
Problem
Existing autonomous vehicle systems often rely on expensive LiDAR technology for localization and perception, increasing costs and complexity. This reliance can limit the accessibility and scalability of autonomous solutions, particularly in controlled, low-speed environments.
Solution
PerceptIn offers a vision-based self-driving platform designed for deployment in low-speed environments like university campuses and industrial parks. The system utilizes a sensor fusion approach, combining visual perception, visual-inertial odometry (VIO), and GPS/VIO fusion to achieve reliable localization without LiDAR. By extracting semantic information through deep learning and spatial depth through computer vision techniques, the platform enables autonomous navigation with enhanced safety and flexibility. The system's modular design allows for integration into various vehicle types, reducing travel costs and time while increasing safety.
Target Audience
The primary target audience includes universities, industrial parks, and local governments seeking to implement autonomous transportation solutions within controlled, low-speed environments.
Features
- Visual perception module combining deep learning and computer vision for semantic and spatial understanding.
- Visual Inertial Odometry (VIO) fusing camera and inertial measurement unit data for real-time position updates.
- GPS/VIO fusion for accurate localization, even in GPS-denied environments.
- Planning and Control module acting as the "brain" of the driving system.
- Computer vision-based sensor fusion for reliable localization without LiDAR.
- Modular design allowing integration into various vehicle types.