Go up one level Go home

1 Complexity of the environment


Highly Automated Vehicles (HAV in effect CAV levels 4/5) will use a number of systems to find their way from one place to another safely, efficiently and causing minimum of disruption to other vehicles. The route planning and finding system will rely on accurate digital maps; safety systems will use multiple sensors and use artificial intelligence to deal with other vehicles, pedestrians, street furniture and unexpected items like incidents and discarded boxes on the road. All of these must be sensed, interpreted and an appropriate action selected under different weather and light conditions; all these influences pose significant challenges. Human drivers are reasonably good at dealing with these issues, at least most of the time. The use of multiple systems is sought as a way to overcome problems when one of them fails, for example, inaccurate digital maps; in such a case perhaps a second-best solution can be explored and found.

HAVs will need an ability to know in advance what to expect from the environment, sense it in real time, update its understanding and interpretation of the real context and select an appropriate action poses significant challenges.

At present, the infrastructure requirements of HAVs are unclear. The performance of HAVs is dependent upon the complexity of the environment in which they operate. On the one hand, providing assistance to HAVs through infrastructure adaptation could ensure that they perform more reliably, but on the other, HAVs could develop to a high standard of driving without the need for costly and time-consuming infrastructure adaptations. Without infrastructure adaptation, HAVs are more scalable and could be with us much sooner, but they may not work as well.

There is clear evidence that AVs can work reliably with simplified operating environments. For example, the Heathrow Pod is an AV that has been operating since 2011 on its own pathway at London’s Heathrow Airport. In the Netherlands, the automated Parkshuttle connects the Kralingse Zoom subway station with the business park Rivium in the city of Capelle aan den IJssel, with its own segregated roadway. There will be many pilot schemes for services of this nature, for example, the Greenwich Automated Transport Environment (GATEway), expanding to public mixed roads.


2 Digital Mapping and Connectivity

(See also Connectivity)

HAVs will make use of digital mapping. They will constantly compare information seen by their sensors to information stored in the map. If the information stored in the map is out of date, the vehicle may be unable to find an efficient way to its destination and may need to either stop or relinquish control to the human driver. The road layout could be altered as a result of roadworks, road collisions, disabled vehicles or other hazards that are less detectable, such as oil on the road. These situations are known to present a challenge to AV technology. Rather than relying on AVs to correctly sense, interpret and respond to such incidents, which could fail and lead to severe consequences, it may be better to communicate the location of such dangers to AVs in advance.

Ways to tell the AV about roadworks or closures are being investigated. These might include updates to digital mapping and/or infrastructure to vehicle communication solutions. Real-time communication of when roadworks are beginning and ending would be useful. This would require new methods of working across the road maintenance industry, and mechanisms to ensure that the information is trustworthy from point of origin all the way to use by the AV control system. At present, connected vehicle technology is being developed as a driver assistance system that notifies the driver of an incident. If information is not successfully delivered, consequences might, therefore, be high risk. Technical communities in vehicle connectivity and vehicle automation working more closely together should be able to find solutions that serve them both.

3 Road Markings

Some AV technologies, particularly lane-keeping systems, rely on clear and consistent road markings. Significant deterioration or unusual use of road markings may confuse AVs or even lead to an incident. The importance of line markings and signage for CAVs is recognised by some organisations. The European Road Assessment Programme (ERAP) and the European Car Assessment Programme (ECAP) released two papers:

  1. 1. Roads that Cars can Read - A Consultation Paper, June 2011
  2. 2. Roads that Cars can Read - A Quality Standard for Road Markings and Traffic Signs on Major Rural Roads, November 2013

Limitations for lane-support systems were identified in the consultation report as follows:
“Currently, the main limitations identified for lane departure systems because they rely on greyscale images are, other than mud, heavy rain, fog and snow:
  • Old road markings not completely obscured even if blacked out
  • Bitumen lines used to seal cabling or drainage in the roadway
  • Faded indistinct lines on asphalt surfaces
  • Slightly faded lines on concrete road surfaces that present poor contrast
  • Lane markings not in normal use
  • Discontinuous markings.”

The 2013 paper calls for an “establishment of an intervention and maintenance policy to ensure that road markings on Europe’s roads remain visible to the driver and the intelligent vehicle at all times, irrespective of weather conditions”. The document endorses the European Road Federation’s (ERF’s) definition of a good line marking, whose “minimum performance level under dry conditions is 150 mcd/lux/m² and which has a minimum width of 150 mm for all roads; for wet conditions, the minimum performance level should be 35 mcd/lux/m²”.

4 Safe Harbour Areas

In full motorway/highway pilot mode, vehicles will be travelling at high speeds with the human driver disengaged from the driving task. The driver may not be ready to regain control of the vehicle before it reaches the end of its operational envelope. This could be for several reasons, such as:

  • The driver falls asleep, suffers a debilitating incident (e.g. heart attack) or becomes otherwise distracted
  • CAV system malfunction or mechanical problem
  • Deterioration of environmental conditions
  • Detection of incident ahead, such as disabled vehicles in the carriageway, which CAV is unable to negotiate

In this situation, the vehicle will need a safe area to stop and wait for the driver to be ready, or for conditions to improve to the extent that the automated control system is able to proceed.

On UK and other countries’ motorways, a continuous hard shoulder has traditionally been provided for so that vehicles can pull over if necessary. These are sometimes being replaced with 'All Lane Running', where the hard shoulder is permanently or partially converted to a running lane combined with emergency refuge areas at intervals of 2.5km. Some motorways have been converted to ‘Dynamic Hard Shoulder Running’, which involves retaining the solid white line to indicate the presence of a hard shoulder, but opening the hard shoulder to general traffic during busy times via indications on overheard signage.

Non-motorway roads (including high-speed single and dual carriageway roads) have no requirement for safe harbour areas to be provided. On these types of roads, there may be a need to increase confidence that HAVs can reliably handle a greater range of scenarios.

Research may be needed into the most appropriate form of safe harbour for CAVs. The advantage of a continuous hard shoulder is that there is always somewhere to stop at short notice. A disadvantage is that a hard shoulder is not necessarily a safe place to stop. Vehicles travelling in the nearside lane of the motorway can veer into the hard shoulder if the driver loses concentration. Research by the UK’s AA, a national breakdown cover service, indicates that, on average, 836 people in the UK have been killed or injured each year in incidents on the hard shoulder and in lay-bys.

For roads where formal, traditional safe harbour areas cannot be provided, it may be possible to consider informal stopping places, such as low-speed side roads or service stations, ensuring that AVs can navigate their way to these in case of emergency.

5 Junctions

Merge and diverge sections on freeways and motorways and urban road junctions are the most significant conflict points for traffic flow. HAVs will be able to perform legal manoeuvres at these junctions. They may even be able to respond reasonably to heavy congestion where some manoeuvres require courtesy from drivers. Nevertheless, the capability to communicate between vehicles and between vehicles and infrastructure is likely to improve performance. For example, communications between vehicles and traffic lights will enable better signal timings. It has even been argued that when all vehicles are CAVs they will be able to negotiate among themselves to agree in what order to cross the junction. It has long been argued that better communications with other vehicles and infrastructure can achieve much better performance than existing traffic lights (http://jair.org/media/2502/live-2502-3761-jair.pdf).

A recent study by TRL in the UK looked into how human drivers will respond to HAVs in the presence of junctions. It used a driver simulator to explore how human drivers would adapt to the presence of AVs at junctions and when overtaking. The study found that a degree of adaptation was observed; for example, human drivers accepted slightly shorter gaps when intercepting an AV but the difference was very small.

6 Auto-Valet Parking

The idea of ‘autonomous valet parking’ has been discussed for many years, and several car companies have demonstrated systems that show an AV capable of searching for, detecting and manoeuvring into a parking space with no human intervention. This creates opportunities both for the user and for the infrastructure provider. Firstly, for the user, this would enable a driver to park somewhere close to the entrance of a car park in a designated vehicle drop-off area and continue directly to their destination without the time and stress of parking a car. The car could be subsequently summoned to a collection area.

For the infrastructure provider, there could be an opportunity to significantly increase car parking provision within a given land area. In principle, empty HAVs could park themselves very efficiently, without the need for human occupants to open the doors. This alone could enable 20% more spaces to be provided within a car park. AVs could also block each other in and let each other out when necessary. A study by Audi suggested 2.5 times the number of vehicles could fit into a car park using this method compared to human-controlled vehicles.

A video was created that simulates how AVs could park themselves in a highly efficient manner. The idea is that each vehicle can be blocked in by up to two other vehicles. If a blocked vehicle needs to exit, the other vehicles will move out of the way to allow passage.

Enabling vehicles to manoeuvre as directed by the car parking system could be challenging. Some form of remote control access will need to be granted to the car park operator. Safeguards would also be needed in the event that a vehicle does not respond, and how to retrieve any vehicles that may be blocked.

7 On Street Parking (and narrow roads)

Parking at the side of the street often results in insufficient road width for two-way traffic. Drivers decide amongst themselves who will go first, which is often communicated by a hand gesture or a flash of the headlights (a form of vehicle-to-vehicle communication amongst drivers). HAVs are expected to struggle with this form of traffic arbitration. If HAVs are going to negotiate narrow streets with significant levels of on-street parking, it might be necessary to consider options, such as the removal of on-street parking so that two vehicles can pass each other, or conversion of streets to one-way operation.

8 Parking Demand

Parking demand will significantly impact the way in which we plan, design and build car parks in years to come. Predicting the effect that the introduction of HAVs may have on parking demand is difficult, particularly in the short-to-medium term.

In the long-term, when HAVs become commonplace, most experts agree that although the overall number of journeys may increase, demand for parking in central areas will decrease. Research from Eno Center for Transportation indicates that by replacing personal vehicles with AVs, parking costs will significantly reduce. Other studies, such as ‘Potential Impact of Self-Driving Vehicles on Household Vehicle Demand and Usage' by the University of Michigan predict that AVs serving multiple residents within a household could reduce vehicle ownership by up to 43%, but, at the same time, increase travel per vehicle by up to 75%. Reduced parking demand, and increased mileage, would be a result of HAVs either continuously circulating to pick up and drop off passengers, or driving themselves to a car park in a non-central area to park. Cars could be aggregated into large car parks located away from central areas.

Space that is freed up from central parking areas and suburban streets could be considered when planning land use for the future. The future widespread use of AVs could make much of this space available for redevelopment.

Given that the transition to less parking demand in central areas may not occur for decades, an adaptable approach to car park design could be considered. A car park vision produced by US Architect firm Arrowstreet shows a building structure that was previously used as car parking space converted for alternative uses, such as residential, offices, recreation and a rooftop garden (click the right arrow to see 'phase 2'). They even suggest an area to accept delivery by unmanned aerial vehicles. The important point is that the multi-storey car park has floor heights that enable the structure to be adapted to other purposes as car-parking demand decreases over time.