Automated-Driving Control Unit

A complete GPS module for real-time communication between the control center and cloud is one of the major requirements for autonomous drive. Therefore, the onboard computing unit shall support LTE/5G/V2X wireless network connectivity for real-time data streaming, event notification, and communications so that the cloud center can always receive information and updates from the vehicles that are connected to.

The network must be able to transmit huge amounts of data at very high speed, with low latency, in all conditions (weather, traffic state, etc.) and without interferences. Additionally, it must be safe against hackers or terrorist attacks, and it must be able to work to some extent even in failure conditions.

While GPS is an essential function for autonomous vehicles, it's not sufficient by itself. The GPS signal is blocked by canyons, tunnels, radio interference, and many other factors, and these outages can last for many minutes and longer. To supplement the GPS, the autonomous vehicle uses inertial guidance with the combination of accelerometers and gyroscope, etc. These sensors provide data on the rotational and linear motion of the platform, which then is used to calculate the motion and position of the vehicle regardless of speed or any sort of signal obstruction.

Sensors are a key component to making a vehicle driverless. Camera, radar, ultrasonic and LiDAR enable an autonomous vehicle to visualize its surroundings and detect objects. Cars today are fitted with a growing number of environmental sensors that perform a multitude of tasks. However, each sensor alone has its limitations.

The sensor perception task includes three parts: localization, detection and tracking, all of them achieved through data fusion performed at different levels.

Localization is usually performed by algorithms that fuse data from GPS, IMU, and LiDAR, resulting in a high-resolution group map. Vision-based deep -learning technologies are achieving accurate results for object detection, as they can autonomously handle huge amounts of data. Deep Learning techniques have also demonstrated their suitability for object tracking relative to approaches based on computer vision.

Decision-taking is one of the most challenging tasks that AVs must perform, especially in awkward situations. It encompasses prediction, path planning, and obstacle avoidance, all of them performed on the basis of previous perceptions. And finally, the HMI is to provide information about driving.

The computing performance of the computation unit is crucial as it involves: sensor data collection and interpretation, sensor fusion, virtual world model updating, AI action planning and car control command issuance. Meanwhile, it is crucial to keep lower power consumption and enhance memory capacity on the same platform.

Safety has long been a key topic in the automotive industry and the list of requirements continues to grow. ISO26262 is the new automotive functional safety standard for passenger vehicle industry and governs the development of safety-related electrical and/or electronic (E/E) systems within road vehicles. ISO 26262 imposes stringent requirements that encompass the entire life cycle of a system, from concept phase to development, production, and decommissioning. It addresses the overall safety management process and covers relations with suppliers and interfaces for distributed development.

The government has a role in testing and deployment to guide vehicles with automated driving systems to have profound effects on society by increasing safety and mobility. Challenges for government include ensuring the systems are safe, determining who is liable for accidents, and public interest in adopting the technology.

Technological improvements in the infrastructure will also be necessary. First, those aimed at helping AVs to perform the perception tasks: horizontal and vertical road signs must be clear and complete, road layouts should be as smooth as possible, etc. Second, V2I-related technologies must be deployed.

The GIGABYTE PILOT deploys powerful main processor and GPU solutions to use the power of AI and deep learning to deliver an autonomous driving solution from data collection, model training, and testing in simulation to the deployment of smart, safe auto-pilot cars.

The GIGABYTE PILOT controller can process the vision sensor data, from multiple cameras, LiDARs, radars, and ultrasonic radars and offers a broad portfolio of high-performance automotive solutions for sensor fusion applications. GIGABYTE PILOT creates a comprehensive environmental model by fusing various sensors in and around the car to computes the best driving strategy for maximum driver safety and convenience. It makes decisions and triggers actuators and is designed to meet the highest safety and security standards.

GIGABYTE is planning to assist the overall system design in attaining the desired A-SIL (according to ISO26262) level for safety systems with high efficiency.