Autonomy - Blees

Autonomy Level 4, closer than you think

Perception enabled technology

Technology behind

GNSS

Receivers for GNSS satellite navigation with increased precision thanks to RTK technology enable precise positioning and orientation of the vehicle in the global coordinate system. These receivers use the most popular GPS system, among others.

IMU

Inertial sensors for measuring the forces acting on the vehicle. Thanks to them, it is possible to navigate in the event of a temporary loss of the GPS signal or obtaining an inaccurate location from the lidar.

Odometry

Based on the relative vehicle displacement information obtained from the wheel rotation reading, the system is able to optimize the vehicle positioning method.
Odometry works regardless of environmental conditions and is a good source of position change when measurements from more accurate sensors are not available.

Radars

Like lidars, they allow you to detect obstacles and precisely measure their speed. Their accuracy is significantly lower compared to lidars, but they allow for operation in difficult weather conditions - for example during heavy rainfall.

Cameras

They make it possible to observe the surroundings around the vehicle. If necessary, they allow the remote operator to maneuver. On the other hand, machine learning techniques make it possible to detect and classify obstacles around the vehicle.

Lidars

They allow you to build a spatial map of the vehicle's surroundings, thanks to which the vehicle can locate itself, build its perception, and detect obstacles, their shape and distance from them.

High accuracy lidars

The vehicle is equipped with a number of high-resolution solid-state LIDARs. These state-of-the-art LIDARs do not have any mechanical parts and are therefore more reliable and provide better quality data than traditional rotating LIDARs.

Localization system

An advanced positioning system that is accurate down to the centimetre is a key component for autonomous driving. The system combines data from multiple sources like GPS RTK, lidar registration, visual odometry, wheel odometry, and therefore is resilient to a single data source failure.

Cutting-edge perception

The perception algorithms incorporated in our vehicle are tailored to the environments in which it is operating. They are a perfect example of striking a balance between cutting-edge technologies, such as deep learning and traditional approaches utilizing analytical methodology.

All of the manoeuvres are monitored by an independent safety supervisor module which is designed to continuously keep the vehicle safe.

High-definition map

High definition maps for autonomous driving are prepared for each deployment. They provide prior knowledge about the world, which augments sensor information and increases system reliability. Maps are enriched with additional information such as bus stops locations or any zones where the vehicle should reduce its speed in order to make passengers feel safer.

Smooth ride

The autonomy is constantly adjusting various parameters like speed, acceleration, deceleration, centrifugal acceleration to keep passengers comfortable.

It can also offset minor vehicle issues, like low air pressure in a tire which can be increased later on. All of that is done without compromising the safety of users and others around the vehicle.

Tracking