Intelligence (AI) and Machine Learning (ML) technology is undoubtedly the best path to efficient and precise autonomous driving systems but also one of the most challenging. The traditional in-vehicle technology doesn't have the necessary capacity to perform high processing workloads, especially intensive AI and ML decision-making.
The real brains of most AI and ML technology lie in powerful massive data centers in the cloud— not in an in-vehicle computer. If a moving self-driving vehicle wants to leverage AI and ML, it requires exceptionally high-speed bandwidth and wireless range. Traditional broadband technologies such as 3G and 4G cannot fulfill these requirements.
Self-driving vehicles need to communicate their data to the cloud or other nearby vehicles in real-time. Although 5G is one of the essential technologies for autonomous driving systems, as it allows reduced response times, cars still need to be self-reliant when making intelligent decisions, primarily offline. Autonomous vehicles require far from ultra-low latency communications to benefit from cloud computing to make quick decisions on the road.
Lanner is currently involved in multiple autonomous driving system projects. To fulfill requirements and regulations, Lanner is accredited with ISO 26262 certification in the commitment to expanding edge compute platforms to autonomous driving. As of now (2022), Lanner provides AI-powered edge computing platforms to enable autonomous and intelligent driving.
Most of the time, Autonomous driving vehicles require two in-vehicle computing systems. One computer processes a large amount of sensed data and images collected by cameras and sensors. And a second computer to analyze processed image data and make intelligent and quick decisions for the vehicle.
- Pre-processing collected data. Autonomous vehicles have video cameras and a variety of sensors like ultrasonic, LiDAR, and radar to become aware of their surroundings and the internals of the vehicle. This data coming from different vehicle sources must be quickly processed through data aggregation and compression processes. An in-vehicle computer needs to have multiple I/O ports for receiving and sending data.
- Secure network connectivity. The in-vehicle computing solution must remain securely connected to the Internet to upload the pre-processed data to the cloud. In this case, having multiple wireless connections for redundancy and speed is crucial. High-speed connectivity is also vital for continuous deployments of vehicle updates or "push" updates like location, on-road conditions, and vehicle telematics.
- High-performance computing. Autonomous vehicles may generate approximately 1 GB of data every second. Gathering and sending a fraction of that data (for instance, 5 minutes of data) to a cloud-based server for analysis is impractical and quite challenging due to limited bandwidth and latency. Autonomous driving systems shouldn’t always rely on network connectivity and cloud services for their data processing. Self-driving vehicles need real-time data processing to make crucial quick decisions according to their surroundings. In-vehicle edge computing is essential for reducing the need for network connectivity (offline decision-making) and for increasing decision-making accuracy.
In-vehicle Data Collection and High-performance Processing Solution.
Lanner’s rugged in-vehicle series (such as V6S) are great options for data processing and network connectivity. They provide rich and multiple I/O ports with PoE ports, serial COM ports, video output ports, GPS, USB, and expansion for storage and video transcoding. In addition, Lanner’s in-vehicle computers also provide the necessary network connectivity with Gigabit Ethernet ports and expansion modules to bring LTE, 5G, PCIe, and more.
The majority of Lanner’s in-vehicle series are also compliant with the E13 standard and have passed the challenging MIL-STD-810G shock and vibration resistance certifications. These certifications ensure the in-vehicle computer can resist shock and vibration and support wide temperatures.
a.High performance In-vehicle edge computing system.
Real-time AI processing for self-driving vehicles is quite challenging if the processing relies on cloud computing. It can be nearly impossible to send large amounts of data back and forth to the cloud and expect the autonomous driving system to respond in real-time.
To solve this problem, autonomous vehicle manufacturers can leverage edge computing. Bringing computing to the edge network (edge computing) closest to where data is generated is the key to solving the high volume of data transfers, latencies, and security challenges.
The autonomous driving system can leverage Lanner’s edge computing appliances when it comes to high-performance computing for self-driving and execution. A compact 1U form factor appliances such as NCA-5710 provide the necessary computing power (edge computing) to process data within the vehicle (without sending it to the cloud).
b.Ice Lake-based MCU for Autonomous Driving system
NCA-5710 is optimized for performance, enhanced security, firmware addons, and modulized scalability. The appliance is powered by 10th Gen Intel Core mobile and 3rd gen Xeon Scalable server processors (code name Ice Lake-SP) based on the Sunny Cove microarchitecture.
The Xeon Ice Lake microprocessor is built-in with AI. The MCU is suitable for many applications, including high-performance computing, networking, and intelligent edge. The Ice Lake-based MCU brings high-performance AI to the vehicle at scale (with Intel® Deep Learning Boost).
Other NCA-5710 key features:
- Scalable storage. NCA-5710 is equipped with two 2.5” Internal Bays to expand storage.
- IPMI onboard (SKU B & C). NCA-5710 has an onboard Intelligent Platform Management Interface (IPMI).
- Small form factor. NCA-5710’s compact design can be easily installed anywhere within the vehicle.
- Memory capacity. DDR4 2933/2666/2400/2133 MHz Max capacity of 384GB.
- Expansion: Wireless connectivity (WiFi, LTE, and 5G), Storage flexibility (2.5/3.5 HDD/SSD/NVMe) PCIe expansion, GPU Accelerator, Video transcoding (4K video with H.265 support).
The Ice Lake-based MCU is a perfect fit for in-vehicle edge computing required to make vehicles self-drive. Edge computing is changing the way we drive and bringing along other fantastic value to drivers.
- Reduce data. Edge computing helps vehicles process their data in much closer proximity. This data processing on the edge network may include compression, deduplication, and aggregation to ensure a substantial reduction of large volumes of raw data.
- Faster bandwidth and lower latencies. Since no significant amounts of data need to be sent to the cloud back and forth, the network latencies are greatly reduced, and the need for more bandwidth also decreases. The reduction in the need for more bandwidth saves money, while the lower latencies improve the application's response times.
- Security and compliance. Edge computing processes data on the edge network, whether on a nearby geographic data center, MEC, or the vehicle itself. This local (or regional) processing helps with regulation and security assurance.
- Edge AI. Bringing AI to the edge allows latency-critical vehicle monitoring tasks such as object tracking, detection, location awareness, and data privacy. Edge AI can also be achieved if the collected data (on the vehicle) can be processed immediately, so that decision and prediction-making can happen in real-time without needing remote (cloud) resources.
For more information on the Ice Lake-based MCU for autonomous driving system, please contact Lanner’s sales representative.
1U Rackmount Network Appliance for Network Traffic Management and Virtualized Network Security
|CPU||2nd Gen Intel® Xeon® Processor Scalable Family (Skylake-SP/Cascade Lake-SP)|