Overview
Concrete mixer trucks roar through our city streets and neighborhoods; wherever there is a construction site, there are concrete mixer trucks on the road. With a huge mixing drum, concrete mixers are top-heavy, unstable, and have a rollover rate 10 times higher than average. Their astonishing size also creates large blind spots for mixer drivers, posing a safety hazard to smaller vehicle occupants, pedestrians, bicyclists, small animals, and even the mixer operators themselves.
A ready-mix concrete company utilized a vision assistance system to give the driver an "extra eye" to monitor "blind spots". A "blind spot" is an area the driver cannot see, with the naked eye or through mirror reflections. When the driver's seat is on the left side of the cab, the largest "blind spot" is along the right side of the vehicle. The right blind spot is especially dangerous when the truck is turning right because of the difference in turning radius between the inner wheels.
The developer's concept is to install a Neousys NVIDIA based NRU Edge AI computer on board, connected to a GMSL2 vehicle camera installed around the right rear wheel and aimed at the right front wheel. A monitoring display is set up in the cab to show real-time images. If the system detects a foreign object in the targeted "blind spot," it triggers an alarm to alert the driver and highlights the object(s) on the screen.
The ultimate goal of this application is to improve the safety of concrete mixer trucks. While in the event of an accident, it can also provide valuable images to help identify the cause of the accident and help determine responsibility. The system is a combination of a sentry system and a dash cam.
Challenges
- Extremely harsh operating environment: A concrete mixer truck is either hurrying on the road or located at a construction site during operations, which means extraordinary harsh operating environments for the onboard computer and associated devices.
- Image quality in various lighting conditions: A concrete mixer truck is either hurrying on the road or located at a construction site during operations, which means extraordinary harsh operating environments for the onboard computer and associated devices.
Camera placement
Solution
The system integrator chose Neousys' NRU-51V+ edge AI computer, powered by an NVIDIA Jetson Orin NX8GB, to be ideal for their truck blind spot vision project. Featuring GMSL2 connectors to support the use of higher quality automotive cameras, advanced image processing capabilities to improve image quality in dark environments, and high-throughput data transfer capabilities.
Monitor in truck for blind spot vision 
AI-powered vision detection The Advantages of Neousys NRU Rugged Edge AI Computer：
- Ready and rich I/O deployment with ignition power control
Compared to the demo board previously used in the project, Neousys' NRU-51V+ is a highly mature system with a rich I/O deployment that can fully meet their development needs. The NRU computer also comes with Ignition Power Control (IGN), which protects the computer from electrical fluctuations during vehicle startup and shutdowns. - GMSL2 IP67 camera support and ready-to-use driver
Neousys NRU-51V+ supports 4x GMSL2 automotive cameras via FAKRA Z connectors and is pre-installed with GMSL2 driver software, making it very convenient to design a GMSL2-based vision system utilizing an IP67 waterproof GMSL2 camera. - GNSS support and high-speed data transfer
Backed by Neousys' patented ruggedness technologies, the NRU-51V+ offers -25°C to 70°C operation, 8V to 35V DC input with built-in ignition power control, and military-grade ruggedness (MIL-STD-810H compliant) to withstand the shock and vibration of a concrete mixer truck. - Cost-effective AI
The NRU-51V+ is capable of 100 TOPS of AI inference in a 25W power package, making it a cost-effective choice for developing AI algorithms for ROI, MOD and BSD functions for wheel radius difference warning systems.
