AVs: BT and UWB can protect vulnerable road users

While there’s much talk in the media about the dawn of fully autonomous vehicles (AVs) on our streets, we’re a long way from that reality. This includes a lack of development in technologies to protect vulnerable road users (VRUs).

Most of today’s “autonomous” vehicles offer at best what the Society of Automotive Engineers (SAE) calls “driver support,” as opposed to “automated driving.” What this means is that the person in the driver’s seat is the one driving the vehicle at all times, even when the driver-assistance systems engage.

To move into the domain of automated driving (Levels 3–5 in the SAE J3016 Levels of Driving Automation Scale), the vehicle must be capable of taking full control in certain situations. This is a significant step up from current driver-support features.

Visible progress at autonomy Levels 3 and above hasn’t been as fast as it has for driver-assistance technology. One reason is the still-evolving regulatory landscape. Another is the need for the enabling technology to mature to support autonomous driving in a greater number of situations.

(Source: SAE International)

Regulatory foundations

Changes to the Vienna Convention on Road Traffic, which lay the groundwork for higher levels of vehicle autonomy are in the pipeline. Assuming these aren’t rejected, they’ll see the regulatory frameworks to govern AVs emerge over the coming years, particularly around the features related to Levels 3 and 4 in the SAE scale. Indeed, early 2022 saw the Insurance Institute for Highway Safety announce its ratings program for vehicles with partial automation. The greater certainty that results from developments such as this will drive an uptick in the pace of innovation to tackle the technological challenges.

Moving up the autonomy scale

Much of the current R&D taking place around higher levels of automation focuses on motorway driving. Relatively speaking, this is one of the simplest public road environments. There are no complex layouts, such as roundabouts, to navigate. And under normal conditions, there are no people walking, wheeling, or cycling — all classes of VRU — present.

Autonomy in urban areas

Urban areas are altogether different, partly due to the increased complexity of the road layouts and partly because of the number and variety of VRUs in the environment. From an automaker’s perspective, protecting VRUs is extremely challenging. Most VRUs can move in sudden and unpredictable ways and are difficult to detect using the technologies currently employed in vehicles capable of higher levels of automation. Cameras, for example, rely on line of sight and can’t see a pedestrian about to step into the carriageway from behind a parked van.

For vehicles with Level-3-and-above automation capabilities to be considered safe enough to operate autonomously in urban areas, detection of VRUs needs to develop significantly.

No single technology on its own will be sufficient. The key will be to use multiple, complementary technologies. Work is already underway to use LiDAR to assist in-vehicle cameras, for example. One area that hasn’t yet seen significant development in terms of VRU protection, though, is the use of terrestrial radio technologies, notably Bluetooth and ultra-wideband (UWB).

Terrestrial radio technologies like Bluetooth and UWB in AV systems could help protect vulnerable road users. (Source: u-blox)

Bluetooth and UWB: ripe for development

u-blox recently ran a study that demonstrated the promise of using terrestrial radio ranging to complement GNSS for high-precision vehicle positioning in urban areas. A natural extension of this would be to explore its use as a means of protecting VRUs. Bluetooth is already built into virtually every new vehicle, smartphone, and wearable device, and the low cost of components means it can feasibly be integrated into others, such as micromobility vehicles, where it isn’t already. UWB is also gaining momentum, having begun to appear in smartphones and being adopted by automakers for applications like keyless entry.

How terrestrial radio technologies could work

So how would a Bluetooth or UWB VRU-detection system work, and what further R&D needs to happen to enable these technologies to form part of future vehicles’ autonomous driving capabilities?

A Bluetooth- or UWB-enabled detection system would require the vulnerable road user to have some kind of tag on or close to their body — or on the vehicle they’re using. This could be within a smartphone or watch. The tag would send out a signal periodically, enabling nearby AVs to calculate the distance to the VRU and the angle from which the signal is coming. The solution offers enormous promise but needs significant development in a number of areas.

Challenge 1: establishing data trustworthiness

Firstly, how much influence should each piece of data coming from the Bluetooth or UWB ranging system have over the vehicle’s decision-making? Given a VRU could be moving, it’s important the decision-making system understands how recent any data coming from a VRU is, and therefore, how much uncertainty might exist around it. How do you then integrate it with data coming from other systems, which may conflict?

We need to establish robust processes and metrics that assess the integrity of the Bluetooth and UWB ranging information. This will involve developing a much deeper statistical understanding of what the radio signals look like in different situations. Creating these models will enable the systems to spot outliers and append certainty or trustworthiness ratings to each reading they collect. This will help the vehicle’s control systems to assess the influence each data point should have.

Challenge 2: protecting against bad actors

There will unfortunately always be attempts by unfriendly actors to cause collisions by disrupting AVs’ safety systems. It’s essential, therefore, that these critical collision-avoidance capabilities are sufficiently robust to protect those around and inside the vehicle. UWB has an inherent strength in this regard, in that it’s capable of secure range measurement that isn’t susceptible to man-in-the-middle attacks. Confidence in its measured range will therefore be higher than with Bluetooth and other conventional technologies.

That said, whichever technologies are adopted, defending against the actions of bad actors will be essential. Incorporating security measures like authentication, plausibility checks, and misbehavior detection are all parts of the story. Another will be to incorporate the signatures of feared events into the signal models we touched on above, to increase the ability of the vehicle’s control systems to identify when there’s an attack and know how to respond.

The other important aspect of any defense against attempts to disrupt collision-avoidance systems is to use multiple technologies in parallel, as we’re proposing. Organizing an attack on several vehicle systems simultaneously is significantly more complex than targeting just one, so this redundancy increases assurance levels enormously.

Challenge 3: privacy

Anything that tracks individuals’ movements raises questions over privacy — as we saw in the early days of the pandemic when governments were seeking to roll out Covid-19 contact tracing using people’s smartphones. Technology solutions exist to decentralize and anonymize the data, but more will need to be done to protect people’s information and ensure individuals retain control over what they share, with whom, and for what purposes.

Challenge 4: vulnerable road users without tags

While most teenagers and adults always have a smartphone or smartwatch with them, to provide Bluetooth- or UWB-enabled discoverability, there will always be VRUs who don’t, including young children and older people.

This further underlines the criticality of using multiple technologies for collision avoidance. A market may also emerge for personal protective tags: low-power Bluetooth and UWB wearables with the sole purpose of notifying nearby AVs of an individual’s presence.

Challenge 5: ease of use and integration

Lastly, there needs to be further development around both Bluetooth and UWB to focus them on this application space. Key requirements will be to reduce costs, simplify deployment, enhance usability, and enable better integration with other parts of the ecosystem.

Time to innovate

As we’ve outlined here, the autonomous vehicles of the future are going to need sophisticated systems to protect vulnerable road users, and this will require designers to combine technologies in new ways to deliver exceptional levels of safety. Bluetooth and UWB show huge promise when it comes to localized ranging and are already being used in the vehicle domain for other purposes. We therefore see exciting opportunities for both technologies to make tomorrow’s vehicles safer and thereby unlock those higher levels of automated driving.