oToSLAM - Multi-camera Vision-AI Positioning System for Autonomous Navigation

Autonomous robots and unmanned vehicles require precise positioning to navigate efficiently, yet many existing solutions depend on costly HD maps/lidar or GNSS/RTK signals. oToBrite’s innovative multi-camera vision-AI oToSLAM system offers a breakthrough alternative by ensuring reliable mapping and positioning in both indoor and outdoor environments—without the need for external infrastructure.

Utilizing four automotive cameras, an automotive ECU (<10 TOPS), and advanced vision-AI technology, the system integrates object classification, freespace segmentation, semantics/3D feature mapping, and low-bit AI model quantization/pruning. This cost-effective solution delivers high positioning accuracy of up to 1cm (may vary depending on the environment & additional sensors), outperforming conventional methods in both affordability and precision.

Already deployed in real-world applications like Automatic Parking Assist (APA) and Autonomous Valet Parking (AVP), oToBrite’s technology is making an impact to enable XPeng Motors to implement APA, while its GMSL automotive cameras are also integrated into multiple NVIDIA and Intel edge AI platforms. Leading industrial computer providers, including Advantech, Axiomtek, ADLINK, AVerMedia, Vecow, and Neousys, also support its GMSL automotive cameras.By eliminating highly dependency on HD maps and GNSS/RTK signals, oToBrite’s AI-powered SLAM is setting a new standard for autonomous mobility. To learn more, visit: www.otobrite.com.