Augmented Pixels releases the hardware agnostic Autonomous Navigation Platform and fast SLAM for Home and Industrial Robotics
SAN FRANCISCO, June 20, 2017
As IoT and hardware components become simpler and more accessible for the engineering purposes, a lot of big corporations, as well as start-ups, try to create their own robotic platforms for different purposes (e.g. autonomous vacuum cleaners, robots for warehouses, etc.).
The possibility to program a robot for autonomous driving/flying with high (~ 1 cm) accuracy is a necessary and critical competent of such robotic platforms.
Creation of the Autonomous Driving component requires a lot of investment in computer vision, machine learning and hardware optimization R&D with not trivial risks.
Reacting on this demanding problem Augmented Pixels releases hardware agnostic Autonomous Driving and Navigation Platform, based on the proprietary Augmented Pixels’ SLAM technology, optimized to run with very high performance and power efficiency on mobile low-power CPU and basic sensors like mono RGB Camera/ Stereo and IMU.
Minimal sets of sensors:
1. Configuration 1: Mono camera (VGA, FOV > 70, fisheye is preferred) + IMU + CPU similar to Raspberry Pi 3 and higher;
2. Configuration 2: Stereo Camera + IMU + CPU similar to Raspberry Pi 3 and higher;
3. Fusion with accelerometer, gyro, sonars, wheels encoders etc.;
Performance: 47+ FPS on Raspberry Pi 3, 65+ FPS on Odroid C2.
CPU: ARM + SLAM already ported on several specific chips*.
The full list of available features:
1. Visual (camera-based) SLAM: real-time tracking / 6DOF, mapping (point cloud), sensor fusion with IMU, sonars, lidar.
2. Real-time analytics: obstacle detection, scene segmentation, surface reconstruction, etc.
3. Cloud mapping infrastructure (create, save, store, update, merge, and match maps).
4. User interfaces for management and programming of autonomous driving (e.g. way points setup).
I strongly believe that our Autonomous Driving Platform democratizes the market and removes the last significant barrier in mass creation of autonomously programmable robots that could perform very important business functions!