Traffic Engineering - August 2019
Passive pedestrian detection system offers an alternative way to use LiDAR for traffic safety
At the George E. Edgecomb Hillsborough County Courthouse in downtown Tampa, there is one primary mid-block crosswalk for pedestrian access to/from the main parking garage. Lack of attention by drivers has led to numerous vehicle-pedestrian collisions/near misses with pedestrians trying to reach or leave the courthouse.
THEA originally envisioned recruiting participants to download a smartphone application that would use the device’s geolocation to provide safety alerts to nearby vehicles. However, testing found that the geolocation data provided by the smartphone had insufficient precision to support safety applications, thus requiring THEA to look into procuring alternative pedestrian detection equipment.
THEA found that the LiDAR technology met their desired performance requirements, while the radar and video technologies did not.
The LiDAR units installed at the crosswalk operate by collecting high-resolution 3D representations of objects using pulsed laser light. These scans are then analysed to identify pedestrians (as well as bicyclists) within the defined area of the crosswalk. The LiDAR is programmed to scan a small region of the sidewalk to locate pedestrians that may begin to cross. Moving pedestrians are tracked and their predicted trajectory is calculated and used to identify possible collisions based on the vehicle’s location and speed of approach.
When a pedestrian is detected in the crosswalk, the LiDAR creates a personal safety message (PSM) and sends it the Roadside Unit (RSU) that is installed near the crosswalk; the RSU then signs the PSM and broadcasts it to the Onboard Unit (OBU) of the oncoming vehicle if the vehicle’s trajectory indicates a possible collision with a pedestrian in the crosswalk.
This triggers an audiovisual alert within the vehicle, where the passenger vehicle’s rearview mirror displays a symbol depicting a pedestrian in a crosswalk and a warning symbol in bright, primary colours (see Figure 2). Simultaneously, several brief audio tones are emitted within the vehicle. Due to safety issues (e.g., to avoid pedestrians being distracted), the pedestrian is neither notified nor required to take any action to trigger the alert, thus making it a passive detection system. It is the driver’s responsibility to respond to the alert by either decelerating (preferably) or swerving to evade the pedestrian. After several seconds, the visual alert disappears from the rearview mirror.
Comments
There is some optimism for the performance of AV but some doubt about the standalone pedestrian detectors, especially ones that attempt to predict if pedestrians are intending to cross.
Is this a high-tech solution ($100,000 or so) looking for a problem that would not exist if all drivers (and pedestrians) concentrated on the primary task?
Proposed Protected Intersection for British Bicyclists Is 'Genius' Says Chris Boardman
The so-called CYCLOPS junction, which stands for “Cycle Optimised Protected Signals is said to maximize the opportunities for safe cycling—and walking—but does not affect the overall junction performance for motorists.
The CYCLOPS junction is an orbital cycle route separating cyclists from motor traffic. It is claimed that pedestrians also benefit from the design.
The junctions will be rolled out across Greater Manchester’s active-travel-oriented, 1,800-mile Bee Network, with the first to be installed soon in Bolton and Hulme, South Manchester.
Comment
We have looked at the detailed paper by JCT Consultancy (the developers of the LINSIG modelling program) for whom there is significant respect.
Some initial thoughts:
- Looks like there are two marked foot crossings on the same (bottom right) corner. Is there a different function for each?
- It appears the road was originally two lanes of vehicular traffic in each direction but now reduced to one.
- Do pedestrians have to traverse three times as many kerb ramps compared to standard design? (normally 2 but 6 in this design for movement left to right)
- Pedestrian walking paths have offset (dog leg) layouts instead of straight ones.
- The motorist heading from bottom to top has to drive across five crossings at this single intersection.
- The traffic signals appear to be after (not before) the bike crossings.
- Cyclists going from bottom to top have a circuitous path, they may follow the car path instead as it is direct
Further comments
Is this approach very UK-centric? In the UK, for example, it is very rare to allow parallel pedestrian movements with vehicle movements – they use a lot of scramble crossings to avoid vehicle-pedestrian conflicts. Also, in the phasing diagrams from the JCT paper, there were many examples of ‘split approach’ phasing (all movements from one approach running together). This is VERY inefficient and can lead to 5 phases at a simple cross intersection (where just 2 would normally suffice).
In the inner Sydney context especially, where would we find the extra space to provide both cycleway and pedestrian crossing on each approach? And the further apart the stop lines become, the longer the intergreen period required and thus lost time in the signal cycle.
Will Sydney cyclists be prepared to travel the extra distance around the intersection on the one-way track, or give way to pedestrians as required at each of the crossings given red-light running behaviours and aggressive actions on footpaths. The temptation would be to just take the shortest path.
Despite the reservations we would recommend TfNSW (former RMS Network Operations staff) undertake a proper analysis using NSW examples both simple and complex.
UK Department for Transport using AI to conduct biggest ever study of UK’s road markings
The UK Department of Transport (DfT) has awarded £2m (US$2.5m) to develop the most thorough understanding ever of the country’s road markings as part of a wider £350m (US$435m) funding package for improving the quality of local roads. By having analysis on the quality of nearly 100,000 miles (160,000km) of road, the DfT will have a clearer-than-ever picture of where investment is needed and will be able to advise local councils on areas that may need it most, improving road user safety on some of the country’s busiest local roads. A recently published local roads funding report from the Parliamentary Transport Select Committee showed that 11% of all local roads were in a poor condition and a further 25% showed deterioration, with the study concluding that the repairs would need £10m (US$12.4m) to address.
Comment
Can we please have a similar program in NSW? (Don’t get me started.)
Cars have gotten smart. Can roads catch up?
As more and more intelligent cars and autonomous vehicles hit the road, some engineers are thinking about what can be done to smarten up the streets on which they travel.
Doing so could allow smart cars and trucks to exchange information with other vehicles, traffic-management centers, and private companies about traffic congestion, accidents, and weather conditions. The key to making it happen is an Internet of Things system that includes sensors embedded in the roadway and on traffic lights.
Existing intelligent transportation systems provide some of those features. Information gathered by traffic-monitoring cameras, for example, is being used to adjust traffic signals in real time to ease congestion. Sensor-equipped parking lots can notify drivers of empty spots via their smartphone. And there are some IoT pilot projects, like Austria’s Autobahn, which uses Cisco’s Connected Roadways system to link 70,000 sensors and 6,500 traffic cameras to monitor traffic and road conditions.
But no comprehensive IoT-based transportation system has been fully deployed, according to IEEE Fellow Phillip A. Laplante, author of “‘Smarter’ Roads and Highways.” The IEEE Internet of Things Magazine article covers the benefits of IoT-enabled roads, security and privacy concerns, and technical standards that could ensure interoperability. Laplante is a professor of software and systems engineering at Penn State Great Valley, in Malvern, Pa.
Comments
A very big, contentious topic. There has been lots of testing by RMS over the years. SCATS (Sydney Coordinated Actuated Traffic Signals) algorithms depend of particular loop parameters beyond just "detection". Motorway sensing is very different to signal applications. It would be interesting to see in the Jacobs report if many of the benefits are to do with ease and cost of installation rather than operating performance.
Read more
Side road activated speed signs
Side road activated speed signs are shaping as a key weapon in the fight against deaths and serious injuries on rural roads according to ARRB.
One of the first has been introduced in western Victoria on the Glenelg Highway near the intersection of the Dunkeld-Cavendish Road near Dunkeld.
When cars approach the Glenelg Highway from the side road, they trigger sensors which drop the speed on the highway from 100km/h to 70km/h.
It is hoped this technology will reduce the risk of collisions at this intersection and others like it in rural areas, without permanently lowering speeds. The TAC and VicRoads will roll out 30 such signs around Victoria.
Australian Road Research Board transport safety expert Dr Blair Turner told the ABC of the uptake in the new technology: “The reason is they are hugely successful in terms of the safety benefits. You see benefits of up to 80% reduction in fatal and serious crash outcomes.”
Comments
How will the temporary speed limit be enforced? If it is not, will main road drivers soon learn that it’s an ‘optional’ limit and thus ignore it?
Traffic incidents ‘down 30%’ on Indiana toll road
Extreme Networks provided the Indiana Toll Road Concession Company (ITRCC) with a system which it claims has reduced traffic incidents along the US road by 30%.
The company says its system is powered by Smart OmniEdge technology, which provides real-time updates on traffic patterns, enabling safer and more coordinated use of the roadway.
Comment
The report seems to be very jargonistic and full of hype. The question is how has it achieved the claimed reduction?
It seems that company has merely done an upgrade of a telecommunications network on a toll road. The safety and traffic management devices/activities that generated safety and traffic flow improvements are the work of the toll road operator. An improved data/communication system would help things run better compared to an older less integrated system but should the lion's share of the credit go to the road operator, not their telecommunications provider?
See the background
AV safety regulation
In a strong signal of support for the country’s automotive and technology industries, the UK government has announced it is developing a process to support the advanced trials of fully automated vehicles, paving the way for on-road testing without a human driver being present.
In response to feedback from industry, the government has also announced that its world-leading code of practice for testing automated vehicles will be strengthened further to set even clearer expectations for safe and responsible trials.
Comments
The UK has updated its requirements for on-road AV trails. It is quite supportive of trials and has a code or requirements. You don't even need to have a backup driver in the vehicle although one must be readily available.
The question is how would all the traffic conditions an AV must operate safely in, be specified. The answer is that they leave it all to the AV developers. That's a concern.
Will self-certification for AV safety go the same way as certification for building works or aeroplanes?