Skip to main content

Safety

All eyes on the road

Thanks to their expertise in automotive and aviation safety, the initiators have applied three safety levels to the pilot ‘Planzer – Dynamic Micro Hub with LOXO’ project. These ensure that the commercial vehicle fitted with autonomous driving technology adopts superhuman skills in traffic.

1

The smart Formula 1 driver

The LOXOFuser AI module is based on transformer architecture. It was primarily developed for last-mile delivery on public roads to recipients’ front doors. To do this, the module orchestrates state-of-the-art technologies such as laser scanning of objects, cameras and sensors. The LOXO Digital Driver™ Software algorithms analyse almost 1 terrabyte of data per hour. This allows the LOXOFuser to operate with the efficiency and skill of a Formula 1 driver.

2

The attentive co-pilot

The LOXOGuard component plays a supporting role, similar to that of a co-pilot. It uses software in which defined, reproduceable situations occur to validate LOXOFuser’s information and decisions. To ensure that the application learns and assesses as many situations as possible, the project partners are training and testing the system under numerous simulated conditions such as rain, fog, snow and darkness, before conducting tests on public roads.

3

The human supervisor

LOXO Digital Driver™ continuously records and analyses vast amounts of data over 360 degrees. Its detection is highly accurate, be it of an object as small as a coin, a bird or a moving vehicle 100 metres away. For the small number of borderline cases when human intervention is required, the LOXO Digital Driver™ Software project partners provide a trained human supervisor. They monitor every individual vehicle in the fleet and take the wheel remotely if required.

LOXO Digital Driver™ Software

Five good reasons

  • AI-based software and human experience ensure optimum road safety
  • Extremely fast data recording and processing significantly improves navigation skills
  • The system is continuously perfected using context and correlation-based learning
  • Highly accurate detection of people and objects in a full 360 degree view
  • Remote human supervision as the final safety net