Get LiDAR perception capabilities with one click! RoboSense releases LiDAR algorithm for the first time

"Don't underestimate the power of your in-vehicle LiDAR." In April this year, RoboSense announced the "Prometheus" plan, which provides partners with a series of software and hardware-based autonomous driving LiDAR solutions based on the concept of "responsibility, openness, and sharing." The company is committed to promoting the development of autonomous driving technology with partners to explore the field of autonomous driving better and faster.

Over the past six months, we have worked together with our first batch of deep cooperation partners from the fields of automotive OEMs, technology companies, and universities to conduct a large amount of practical testing based on different real driving scenarios, and the stability of each algorithm module has been verified.

Now, RS-LiDAR-Algorithms perception algorithm version 1.0 is officially open to "Prometheus" program partners!

A development kit that gives your self-driving car instant LiDAR perception capabilities

RS-LiDAR-Algorithms perception algorithm SDK is a software development kit specially developed for autonomous driving environment perception, including functional modules such as positioning, curb/drivable area detection, lane marking line detection, obstacle detection, dynamic object tracking, obstacle classification and recognition, etc.

The purpose of developing this SDK is to facilitate autonomous driving project developers to conduct secondary development based on this kit, thereby accelerating the development of autonomous driving technology.

RS-LiDAR-Algorithms six modules

High-precision positioning

High-precision positioning is the basis of autonomous driving environment perception. RS-LiDAR-Algorithms includes a high-precision real-time positioning module, and the positioning accuracy reaches the industry-leading level (≤20cm), which is sufficient to meet the needs of autonomous driving.

Roadside/Drivable Area Detection

The detection of drivable areas is a prerequisite for the path planning of the autonomous driving system. RS-LiDAR-Algorithms includes a roadside/drivable area detection module, including roadside detection and drivable area detection, providing a "pathfinding" function for autonomous driving vehicles.  

Lane marking detection

Lane markings are another important piece of information that an autonomous driving system needs for accurate path planning, in addition to the curb. RS-LiDAR-Algorithms includes a lane marking detection module that can accurately extract lane markings based on the weak intensity difference information of the laser reflection echo signal, including: lane markings, road markings, crosswalks and other traffic sign information.

Obstacle Detection

Obstacle detection is a prerequisite for autonomous vehicles to interact with other road users, and is also a basic requirement for ensuring autonomous driving safety. RS-LiDAR-Algorithms includes an obstacle detection module that can detect and output the precise location, distance, posture, size, shape and other information of multiple obstacles in the surrounding area in real time, helping autonomous vehicles to "see" the surrounding environment clearly so as to decide the next action.

Dynamic object tracking

For self-driving cars, what they need to "focus" on is moving objects. RS-LiDAR-Algorithms includes a dynamic object tracking module, which can estimate and output the motion parameters of multiple dynamic objects in the surrounding area in real time, including speed and direction. Based on the speed information, it can further calculate acceleration, angular velocity and other information to help self-driving cars analyze and predict the driving/action intentions of other moving objects.

Obstacle classification and recognition

Classifying surrounding obstacles helps the autonomous driving system to more accurately estimate the action intentions of surrounding objects, thereby formulating more accurate path planning and control strategies. RS-LiDAR-Algorithms includes obstacle classification and recognition functions, which can distinguish obstacles into different categories such as pedestrians, bicycles, cars, trucks, etc. In order to improve the accuracy of recognition, we have also established an industry-leading LiDAR scene database.

Autonomous driving, we need to help each other

The Prometheus Project is a series of software and hardware integrated autonomous driving lidar solutions provided by RoboSense to partners in the autonomous driving field.

So far, around the Prometheus project, we have released three mass-production-level LiDAR products, coupling platforms and supporting algorithms, and a large number of close comrades have joined the cooperation.

We look forward to more partners who are committed to promoting progress in the field of autonomous driving joining our program and making progress together!