Using SDRs to Prototype and Deploy Vehicular Networking for AVs

Article By : Brendon McHugh

SDR-based testing systems can reconfigure their internal digital structure to test the RF behavior of the embedded firmware of AVs in face of new protocols, frequency allocations, and functionalities.

Self-driving vehicles are one of the most hyped technologies in this century, promising a complete revolution of the way we operate transportation. One of the fundamental building blocks of any self-driving vehicle is the radio system, which drives not only the infrastructure of the data network required for autonomous operation but also testing/validation and prototyping of new protocols and algorithms. In this context, software-defined radios (SDRs) enable novel standards and functionalities to be supported on the fly, with minimal changes in the hardware. SDR-based testing systems can reconfigure their internal digital structure to test the RF behavior of the embedded firmware of autonomous vehicles (AVs) in face of new protocols, frequency allocations, and functionalities, such as requirements for safety and traffic management, multimedia services, and navigation.

In this article, we discuss how a high-end SDR can be used and configured to meet the digital-signal–processing (DSP) requirements of prototyping and deploying AV networks by providing low latency; adjustable, wide instantaneous-bandwidth and frequency-tuning ranges; multiple-input multiple-output (MIMO) transceiver channels; and a high-performance digital back end based on the field-programmable gate array (FPGA). We will also discuss the role of SDR-based intelligent transceivers and cognitive radio in vehicular-IoT (VIoT), vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), and vehicle-to-anything (V2X) communication.

State of the art in AVs and infrastructure

The development of AV infrastructure is the first step toward feasible incorporation of these vehicles in our society. Up until now, most AV research effort was directed toward the development of new sensors, such as radar, LiDAR, and computing vision devices, which are fundamental building blocks of the autonomous car. However, sensors alone are limited to the surroundings of the vehicle. By introducing fast and reliable RF infrastructure for AV communication, each car can collect larger sets of information from other devices and sensors in the surroundings, including other vehicles, roadside units, traffic lights, and even pedestrians. Hence, VIoT has the potential to increase the situational awareness of the AV exponentially, improving its decision-making significantly. Figure 1 shows the basic AV infrastructure.

Basic AV infrastructure architecture
Figure 1: Basic AV infrastructure architecture

Radio communication in AVs can be performed in V2V, V2I, or V2X. V2V connection is intended to exchange information between vehicles and can be performed via data broadcasting of position information to all AVs in the vicinity. That way, each AV not only can be aware of the position and velocity of nearby vehicles but can coordinate strategies with nearby AVs to ensure the best outcomes for all. V2I communication is performed between the vehicle and roadside units (RSUs), which provide backhaul and local services for the AVs. Similar to V2V communication, RSUs broadcast their position for spatial awareness, via either WAVE (Wireless Access in Vehicular Environments) service announcements or GeoNetworking messages.

Finally, V2X connects the on-board unit (OBU) of the vehicle to any nearby IoT device, including other vehicles, infrastructure, pedestrians, and cellphones. There are several standards for V2X deployments, including dedicated short-range communication (DSRC) and ETSI IT, which use the IEEE 802.11p standard for the physical and medium access layers. One of the most popular standards for V2X systems is NG-V2X, which provides several AV-specific functionalities on top of the 5G NR interface, including exchange of sensor data between vehicles, remote driving, advanced driving (for vehicle coordination), and vehicle platooning. An acceleration in the development of V2X technologies is expected, given the U.S. Federal Communications Commission’s decision to assign part of the 5.9-GHz band exclusively to cellular-V2X communications.

SDRs for AVs

To fully understand how SDRs can be applied in AV deployments, let’s discuss the basic structure of these radio systems. The general architecture of an SDR consists of a radio front end (RFE) and a digital back end, working as a transceiver with DSP capabilities and connection to external hosts, including PCs and networks. The RFE is composed of receive (Rx) and transmit (Tx) lines, which provide amplification, filtering, and mixing over a wide tuning range. High-end commercial SDRs can reach up to 3 GHz of instantaneous bandwidth over multiple independent channels that can operate in a MIMO configuration. The digital back end is composed of an FPGA with DSP capabilities, including modulation, demodulation, up-/down-converting, and data packetization over Ethernet links (qSFP+). Furthermore, the FPGA can run application-specific algorithms, communication protocols, and even artificial-intelligence and machine-learning applications. The RFE is connected to the digital back end via analog-to-digital converters and digital-to-analog converters. Figure 2 shows the basic structure of the FPGA and how it connects to the AV network.

Basic diagram of an SDR and AV host system
Figure 2: Basic diagram of an SDR and AV host system

The reconfigurability of the FPGA combined with the wide tuning ranges of high-end RFEs allows off-the-shelf SDRs to be applied in a variety of applications with different size, weight, and power requirements. Moreover, SDR-based systems can be easily updated and upgraded on the fly and without any hardware modification. Therefore, as AV standards and technologies evolve — and, with them, the frequency and throughput requirements — SDRs are likely to stand the test of time and function as a high-end test kit for AV technologies. Hence, it is no surprise that most AV deployments are implementing SDRs in testing and processing.

SDRs provide important tools for performance testing of V2X systems because of their reconfigurable tuning frequencies, ability to implement different AV wireless protocol stacks, and compatibility with open-source software tools, such as GNU Radio. In particular, low-latency SDRs are crucial to ensure fast and reliable V2V communication in real time.

This was demonstrated by researchers from the Shanghai Institute for Advanced Communication and Data Science who developed a prototyping performance analysis tool for V2V communication using an SDR platform with universal software radio peripherals. They modified the LabVIEW LTE framework by changing the resource allocation, signal-processing mechanisms, and frame structure, emulating the LTE-V standard and using the tool to generate numerical results based on measured signals from the SDR. Their results showed the effectiveness of the tool, which they believe will be used to evaluate V2V networks and 5G V2X technologies.

Another important development of SDR-based performance evaluation is the Epochs Reference Application (ERA), where Epochs stands for Efficient Programmability of Cognitive Heterogeneous Systems. ERA is essentially a reference application that models a set of interconnected AVs operating in the same environment, based on DSRC, Robot Operating System (ROS) middleware, and GNU Radio. Each AV in ERA generates local occupancy map grids using on-board sensors, communicating with other vehicles through DSRC. The ERA software suite can be used to develop AV technologies that focus on the collective intelligence of a set of AVs, using swarm awareness to create navigation strategies that transcend the limited information obtained by only one vehicle. Finally, MIMO SDRs can be used to simulate interference on OBU transceivers, which must be exhaustively tested to obtain the required QoS.

SDRs can also be used to drive location-positioning systems in V2X networks. The motivation for location positioning comes from limitations in GNSS technology, which tends to fail in urban canyons, tunnels, and underground parks because of signal reflection (also referred to as multipath effects). One possible solution for the problem is to use on-board SDRs with advanced processing tools to calculate the packet receive time, eliminating degradation from multipath effects by using accurate timing and multipath awareness. SDRs can also support multiple radio technologies at the same time, such as NR V2X, LTE-V2X, and DSRC, operating with frequencies higher than 6 GHz.

Moreover, commercial off-the-shelf SDRs are compatible with IEEE 802.11p and IEEE 1609 standards, so one can develop an AV network from scratch using readily available SDRs.

Finally, an SDR-based AV system can be easily upgraded to the latest technologies, supporting all current and future 3GPP standards without the need for hardware modification.

Conclusion

AVs are close to becoming a reality, but several technological developments are still needed before we see AVs navigating through our daily lives. In particular, wireless networks between vehicles, infrastructures, and IoT devices are crucial to extend the reach of sensor data, provide spatial and functional awareness to vehicles, and allow coordination between different AVs to obtain the best traffic outcome.

In this context, SDRs are the best option for prototyping, testing, and deploying RF technologies in the AV infrastructure because of their high degree of reconfigurability, wide tuning range, low latency, high data throughput, and on-board host interface. Therefore, SDRs can be implemented in virtually any V2X communication scheme. For instance, SDR-based platforms can emulate V2V networks based on the LTE-V standard for performance evaluation or simulate a set of interconnected AVs to develop swarm intelligence through V2V. Moreover, SDRs can perform location positioning by using multipath-aware enhancements, increasing the accuracy for limited environments, such as urban canyons and tunnels. Finally, SDRs can significantly improve the useful life of OBUs and RSUs by allowing continual upgrades with state-of-the-art protocols and standards on the fly and without requiring any hardware modification.

This article was originally published on EE Times Europe.

Brendon McHugh is a field application engineer and technical writer at Per Vices Corp.

 

Subscribe to Newsletter

Leave a comment