Connected vehicles: Road-ready yet?

Recent progress with Dedicated Short Range Communications (DSRC) Notice of Proposed Rule Making (NPRM) brings connected cars or V2X — connectivity between vehicles, infrastructure and all road users — closer to reality than ever before. If all goes well, an NHTSA mandate on DSRC in new light vehicles is expected to start around 2020 as a phase-in plan, with completion around 2025.

Regulations for aftermarket devices are expected to come soon after. The mandate is expected to leave auto OEMs to choose the applications and human-machine interface (HMI). This will be the culmination of more than a decade of technology development and standardization by U.S. Department of Transportation (USDOT), automotive OEMs and other industry partners.

Significance of V2X. According to USDOT, V2X technology can positively impact more than 80% of non-impaired vehicle crash types that result in over 30,000 deaths in the U.S. alone. A report by the Federal Highway Administration to Congress states that V2X technology is ready to be deployed in the near future and is expected to yield significant safety and efficiency benefits.

From a consumer’s perspective, V2X will be a part of a vehicle ADAS (Active Safety Driver Assistance System). Initial systems will provide information only, and these systems are expected to evolve into warning and control capabilities. In a future vehicle, information from multiple sensors including V2X will be combined/fused to generate a view of the surrounding environment. Figure 1 gives an example of such sensors including long- and short-range radar, lidar, cameras and V2X. V2X offers unique advantages over other sensors that depend on direct line-of-sight. Information can be received from vehicles not visible to other sensors, giving a much larger field of view. V2X can transmit information directly from traffic control devices, instead of inferring information from camera observations.

Figure 1. Example of a vehicle sensor configuration.

Figure 2 depicts the sensor fusion screen from an ADAS development platform by Renesas Electronics America. Such a platform offers the flexibility to implement an ADAS using all available sensors, for example blind-spot warning from radar, forward collision warnings from combined radar, camera and V2X, surround object detection from combined radar, lidar, vision and V2X, with information presented via an OEM-specific HMI.

Figure 2. Renesas ADAS development platform.

GNSS role and challenges

V2X is built on the assumption that vehicles, infrastructure elements, and other road users are location-aware and can communicate critical information to others around them. As seen in Figure 3, the system will position all communicating V2X entities with respect to the host vehicle and security interface, which validates all relevant DSRC messages. A control area network (CAN) or a similar interface will be needed for direct access to vehicle information such as brake and turn-light status and odometer. Interfaces to long-range connectivity such as cellular networks and other data sources such as maps may also be included. The system will connect to an HMI to display information, and future systems will likely evolve to vehicle control functions.

Figure 3. Components of a V2X system.

Looking at the components of an over-the-air (OTA) V2X basic safety message (BSM), this includes a UTC-based time marker, WGS84-based position, and an estimated position error — all critical data that primarily depend on GNSS. RTCM-formatted data may also be sent as optional attachments. A BSM-like personal safety message (PSM) is also defined for pedestrians with V2X-enabled devices.

As per current Minimum Performance Requirements (MPR), a UTC time source with better than 1 millisecond accuracy is required in a V2X device. While almost all current prototypes use GNSS as source of time, others, such as NTP, may also be used. Accurate time reference is a critical prerequisite for basic DSRC functionality. MPR requires time-marked position estimates with 2D and elevation accuracy of 1.5 and 3 meters or better (1 sigma) under open-sky conditions. The automotive industry has opted to define open sky as unobstructed sky view above 5-degree elevation with seven or more satellites visible with HDOP and VDOP limits. The industry expectation is to use this criteria to select GNSS devices that could eventually support lane-level applications (better than 1.5-meter accuracy).

MPR does not put any requirements on the accuracy of the position error estimate in the BSM. It does require that a vehicle stop transmitting BSM whenever the aforementioned time and position accuracy requirements are not met. This implies that a V2X-enabled vehicle may disappear from the V2X view of others in a dense urban canyon or similar environments, leaving at least two questions for system designers from a GNSS perspective alone. First, how to reliably declare that the system cannot meet time and position accuracy requirements, and second, how to deal with the vehicle itself and other V2X entities that may cease to function or broadcast due to GNSS or other limitations. V2X systems are assumed to include inertial and vehicle sensor integration.

Road Ahead. Starting in 2017, connected vehicle pilots (CVP) in New York, Tampa, Florida, and Wyoming will be the next major milestone for V2X. These deployments will be limited to commercial fleets (taxis, public transit, city/road crews and delivery trucks) and some limited road-user categories.

Among the automotive OEMs, Toyota was the first to offer V2X-based driver-assistance technology as ITS Connect in Japan in 2015. General Motors is the first to announce a V2X technology offering in a passenger vehicle in the U.S. with an initial rollout in select 2017 models. The first phase of V2X deployments will only provide driver assistance information while subsequent iterations are expected to bring in safety-focused functions leading to control capabilities.

There is a growing interest in the cellular industry to support V2X-like communication in an upcoming release of the 3GPP standards commonly referenced as 5G. This would enable low latency, peer-to-peer communication with the advantage of an existing device provisioning/authentication infrastructure, something that needs to be built up for DSRC. However, 5G is still a concept, and judging by the lifecycle of LTE, a 5G deployment will take several years to start and several more years to fully deploy while still leaving some rural areas with legacy technology. A framework to manage commercial traffic vs. likely free safety traffic will also be required. These raise the question as to how 5G alone can support vehicle safety applications nationwide.

The FCC has recently proposed a rule to potentially open up the DSRC band for unlicensed Wi-Fi devices, provided Wi-Fi users do not interfere with the primary safety use. Automotive and wireless industry and other stakeholders are investigating the feasibility of possible co-existence in the future. Among the proposed solutions are the rechannelization of DSRC to use a smaller bandwidth and a mechanism for Wi-Fi devices to Detect-and-Vacate the DSRC band when a safety user is detected.

From a technology point of view, V2X has reached a significant milestone with R&D in various technology areas converging and critical standards being adopted recently. With Toyota V2X offering in Japan and GM V2X commitment in the U.S., customers will have V2X as an option this year, further proof that V2X will be on the roads soon. However, significant further work is needed to address the GNSS accuracy and reliability needed for next-generation systems and to address GNSS-specific vulnerabilities such as jamming or spoofing. The New York CVP, which includes deep urban canyons, will probably be a great opportunity for GNSS and V2X communicates to work together on some of these limitations.