Weather Updates For AVs Could Use Coms Cocktail

Weather updates are going to be crucial to allow autonomous vehicles to adjust to changing road conditions in the same way that a human driver should adjust their speed during heavy rainfall or snowfall.

This kind of information should also be able to inform connected and semi-autonomous vehicles that still have a driver to control them. However, the communications networks required to relay this data are often lacking – even 4G coverage is still quite patchy. This brings forth the claim that 5G connectivity will be worse.

This technology will require even more aerials than are presently available. So, what’s the answer? Well it’s now suggested that dedicated short-range communications (DSRC) might help to fill the void, to handle vehicle-to-everything (V2X) communications to deliver updates. The bonus is that DSRC is supported by the latest European Union directives – as well as by the two of the biggest car manufacturers, such as Volkswagen and Toyota.

5G no worse

Yet, despite the view that 5G might be worse than 4G coverage, Michael Venus, product manager for connected vehicle solutions at Siemens Mobility, says: “We don’t think 5G will be “worse”.  Rather, 5G uses higher frequencies than 4G which have a much smaller range and are also much more sensitive to obstacles like trees, etc. This results in 5G cells being much smaller than 4G cells with an estimated range of 700 meters vs several kilometers.”

“In other words, considerably more base stations will be needed to provide similar coverage (to 4G cells). The actual location of the base station will also be very important in order to get a connection at crucial spots like dangerous road locations. The positions of cellular base stations are usually not chosen based on road safety needs because cellular operators usually want to provide the best coverage for the most people in the first place. Since 4G cells cover a much larger area this is not as critical but for 5G it will be.”

Going hyper-local

Whatever the solution, the future of the autonomous car will necessitate having access to highly detailed, hyper-local road condition and weather information. Petri Marjava, sales director at Foreca explains: “In addition to providing information on current road weather and driving conditions, we are also providing predictive road condition information. In other words, we also transmit a road condition forecast to the vehicle containing information on the upcoming conditions along a given route.

“The length of such predictions can be anything from short-term i.e. next 5-30 minutes to several hours into the future. So, in a situation where connectivity is lost, the vehicle systems would still have forecast information at its disposal. Then again when connectivity is re-established the data is refreshed with latest observations, new forecast etc.”

DSRC: the answer?

To what extent are DSRC the answer to revolve any patchiness in 5G coverage – particularly for handling V2X to deliver updates, such as the weather and road conditions?

Sandeep Sovani, director, global automotive industry at ANSYS responds: “The key point here is about network latency. The choice of communication system depends on the speed and reliability with which specific information needs to be communicated. Some information needs to be communicated rapidly, within milliseconds, whereas a delay of a few seconds is fine for some other information. For instance, an autonomous vehicle that is cruising at highway speed, needs to know about the condition of the road immediately ahead, with millisecond latency else the information is of no use. It is of no use to know of a slippery ice patch after having run over it.

“However, it is fine if there is a delay of a few seconds in communicating some other types of information such as local weather conditions, because these conditions do not change in a matter of seconds. Vehicle driving down the road will record information about road conditions that they encounter – such as ice patches or pools of standing water – and transmit it directly to trailing vehicles via Vehicle-to-Vehicle (V2V) communication. This communication needs to be superfast, with millisecond latency, as it needs to happen before the trailing vehicle hits the same ice patch.”

DSRC effectiveness

“This is where DSRC or vehicle-to-vehicle LTE communication can be effective owing to its low latency. However, DSRC is limited by its inherently short range and this is where cellular networks come in. Cellular networks have a high latency but very long range and are ideal to communicate local weather conditions to autonomous vehicles. Cellular networks will also play an important role in aggregating road condition information from several different vehicles – for instance, location of an ice patch that several vehicles encountered – and broadcasting them to all the vehicles in the area.”

Layers of safety

Venus emphasizes that the addition of connectivity to a vehicle offers more layers of safety beyond the vehicle sensors and even beyond the driver’s field of view. “Providing the driver or the autonomous vehicle with advanced warning is very crucial when it comes to accident avoidance,” he says before adding that even a second can make a difference between an accident and one going about their day safely without incident. He also stresses that the faster the vehicle goes, the faster the need for real-time information.

He adds: “Infrastructure can play a major role in increasing safety in that sense. A vehicle on its own can only detect what is in its direct vicinity using its on-board sensors. With infrastructure support like sensors and DSRC communication, information about obstacles or hazardous conditions like ice on the road can be detected and sent to the vehicles even before the vehicle itself or the driver is able to see the actual location. This can help to save the extra second that is needed to avoid a crash.”

Better decision-making

On-board sensors also help vehicles to understand their surroundings and with predictive road condition information, Marjava says and the automated driving functions can be increased. The vehicle behavior can be adapted too. “To mention but a few examples: one can choose a route with more favorable road weather conditions to enable automated driving (for a larger part of the trip).

“In some situations, such route might be somewhat longer but you will have the choice to make depending on what is preferred at each time: a more relaxing journey with automated driving enabled for the most part of the trip or a quicker one with manual control.” It could also enable drivers to make decisions about what to do during challenging weather conditions, including whether or not autonomous driving mode should be engaged or disengaged. Yet, autonomous driving is likely to make human drivers and their skills obsolete in the future.

He adds: “Automated vehicle can adapt its speed more ‘naturally’ when approaching an area where road is wet for example well before any of the on-board sensors would be able to detect this. In this case ride will be much smoother without harsh braking or sudden need for taking manual control.”

Making the transition

The question that remains is about how the transition can be made from manual driving to autonomous driving. Marjava concludes that it’s important to make this period as safe as possible. “If and when people start to use more and more automated driving (e.g. highway autopilots, robo-taxis etc.), the required skills to be able to handle the most challenging driving situations will deteriorate.”

He thinks there needs to be some consideration about “how can we cope with this paradox and avoid even worse traffic and accident numbers caused by severe weather which is something which will remain beyond human control tomorrow as well”.

Ensuring interoperability

In summation Venus says: “We believe that, especially for autonomous driving, all data that is available and can make the driving safer should be used for that purpose. This especially applies to the data originating from roadway infrastructure, like information about the current signal states on an intersection or pedestrian detection data and weather information. All this information must be made available in an interoperable, reliable way and DSRC has been developed exactly for that purpose and is the only mature technology available today.”


Leave a comment

Your email address will not be published. Required fields are marked *