Autonomous
Software Solutions
Kick start development projects with highly intuitive autonomous functionality. Dataspeed’s software offering enables key self-driving features, such as low-level steering control and sensor object fusions.
Functions
Speed & Steering Control
Dataspeed By-Wire Kit
None
The speed control function applies closed loop control systems to generate throttle, brake, and shifting commands to regulate an incoming speed setting. Additional features are included to make smooth transitions between large speed input changes.
The steering control function allows the user to specify a desired yaw rate or turning radius instead of having to compute a steering wheel angle. Additionally, safety measures are included that limit the maximum angle and rotation rate to stay within user-specified lateral and angular acceleration limits.
These two control functions are combined as the Universal Lat/Lon Controller (ULC), and is bundled with the existing Dataspeed By-Wire Kit firmware.
Lidar Object Tracking
This module processes raw point clouds from 3D lidar sensors to detect objects around the vehicle and track their motion over time. The output contains a list of object metadata with objects’ positions, relative velocities, and bounding box sizes for consumption by higher-level systems.
None
3D Lidar (Ouster, Velodyne, or Cepton)
* Currently under final development, but available for quoting.
Multi-Sensor Object Fusion
This module combines object metadata (position, velocity, size) from multiple sources and merges observations of the same object into a single object track. Various sensor types are weighted differently depending on their strengths and weaknesses to provide an overall more robust object tracking system than is possible with just a single sensor.
Lidar Object Tracking
A combination of 3D Lidar, Radar, and/or MobilEye
* Currently under final development, but available for quoting.
Adaptive Cruise Control
This module inputs an object metadata list from either the Multi-Sensor Object Fusion or Lidar Object Tracking modules, determines if there is another vehicle in the ego vehicle’s path, and adjusts a speed command to regulate the following distance and relative velocity.
On roads where the speed limit is greater than 30 mph, fused lidar and radar data is greatly preferable to a single-sensor source. The radar can detect vehicles at longer range than lidar and ensure there is enough time to react if the relative velocity is large. At short range and at low speeds however, the lidar provides much more useful object tracking capability. Therefore, 3D lidar is always required for ACC to function, but radar is needed if the intended application is to operate above 30 mph in environments with other moving vehicles.
Speed & Steering Control
Multi-Sensor Object Fusion
Long-range radar for high-speed (>30 mph) applications
3D Lidar and/or MobilEye
* Currently under final development, but available for quoting.
Trajectory Following
Speed & Steering Control
None
This module inputs a trajectory consisting of a sequence of waypoints with corresponding speed references and generates desired yaw rate and speed commands to follow the trajectory. The controller monitors the incoming trajectory and automatically slows down before entering sharp turns according to maximum lateral acceleration constraints.
* Currently under final development, but available for quoting.
Stop & Go Control
This module is a special application of ACC that is responsible for driving in heavy traffic or at a red light. The controller would operate at very low speeds to regulate the distance to another vehicle that rolls forward and stops repeatedly.
Speed & Steering Control
Multi-Sensor Object Fusion
Long-range radar for high-speed (>30 mph) applications
3D Lidar and/or MobilEye
* Currently under final development, but available for quoting.
Lane Centering
This module is intended to fully control the vehicle in situations where there are no intersections and minimal chance of objects crossing the vehicle’s path. This would apply to highway driving but could also be used on surface streets between intersections. A MobilEye sensor is required for this to detect lane markings effectively. If coupled with an HD map with the desired route loaded at runtime, RTK GPS can be used to localize and control to the lane center as well.
Lidar Object Tracking
A combination of 3D Lidar, Radar, and/or MobilEye
* Currently under final development, but available for quoting.
Breadcrumb Following
This function allows for recording and playback of a manually driven route. For outdoor applications, this requires RTK GPS to accurately and repeatably follow the recorded path. For indoor applications or other areas with plenty of 3D features, a 3D lidar can be used to localize with similar accuracy using a SLAM algorithm.
3D Lidar is required to detect and stop for obstructions in the recorded path. If multiple lidars are used to expand coverage, the Multi-Sensor Object Fusion function is required to combine the outputs, otherwise just the Lidar Object Tracking function is sufficient.
Trajectory Following
Multi-Sensor Object Fusion
Lidar Object Tracking
RTK GPS (outdoor applications)
3D Lidar
Radar
