City transport planners and transport researchers around the world agree that road pricing is the best way to solve traffic congestion. Chargeable congestion zones around some city centres like London are good, but not great. They are coarse, and only cover a fraction of the busy road network.
Pricing roads for their cost of use is relatively simple, but for one key problem: Tracking the vehicles.
GPS appears an obvious solution for ‘detecting position’ of cars using main roads at busy times, but it is quite unsuitable for many reasons.
NeoMatrix Ltd has developed a new AI method to track vehicles for road pricing.
While vehicles can be monitored remotely, this is highly problematic and costly, and covered in subsequent sections below. The ideal replacement for a consumer vehicle tracking (in-vehicle) for mass monitoring by TfL would have these preferable characteristics:
● Very long-life battery, or self-powered via solar and/or RFID (As in approximately equivalent to current in-vehicle windscreen mounted toll transponders)
● Antenna placement more robust than a GPS or RFID antenna, not requiring windscreen placement
● Supports years long tracking of full road network movements, with independent power source
● More robust and secure than GPS
● Cheaper hardware than existing in-vehicle tracking systems, closer to current in-car toll transponders in hardware costs
● Significantly reduced bandwidth requirements for communicating the daily / lifetime route history of the nearly 40 million vehicles registered in the UK.
● Significantly enhanced user privacy technology compared to legacy tracking systems, while still fully supporting high resolution route tracking for road pricing schemes. Sufficient to have wide community acceptance of the technology as a ‘required’ scheme for all drivers on main roads.
www.neomatrix.uk has developed a novel tracking technology method that has been designed for road pricing. It addresses the problems with GPS listed above, while ‘also’ providing additional benefits, specifically including privacy. Consumer adoption of road pricing schemes will obviously face the hurdle of ‘privacy’ which is as much a barrier to overcome as the technology to mass monitor vehicles itself.
GPS is unsuitable for mass tracking of cars, necessary for an effective road pricing scheme. Such schemes have been touted as the answer to congestion by many countries. GPS has been around for years, but never used in this context, for many reasons as below:
Power: GPS draws high power to ‘hear’ satellites & process orbit calculations. GPS often runs phones flat quickly with constant location use. Congestion tracking can’t have high-power consumption. Devices are expected to operate self-powered for years. Hard-wiring power means retrofitting millions of vehicles at untenable expense, and can be easily defeated by severing power lines if GPS was used. Mass adoption requires secure self-contained power, like current (legacy) toll transponders mounted on windscreens.
Note: Security tracking is very different to congestion tracking. For security, the device may detect & transmit positions each ten minutes to be effective for its security purpose. Road pricing needs a much higher resolution tracking (billing by km), which consumes far more power than a GPS theft tracker.
Antenna: GPS tracking requires an antenna mounted with line-of-sight to the sky. They mount inside the dashboard or intrusively on windscreens to hear signals. Antennas are an Achilles heel. Drivers can sever the antenna wire, or shield with foil to make the vehicle appear not moving, and thus will not incur congestion road use fees. Even if movement itself is detected via sensors, lack of GPS positions means movement is assumed non-main road use, such as on private property or local roads, so not billed.
Heavy earthmoving equipment can’t use GPS easily due to difficulty placing sensitive antennas. Thieves break antennas or wires to steal vehicles, as GPS antennas can’t be fitted easily or securely internally in a solid metal vehicle due to faint signals being blocked. Such vehicles use wireless or GSM tracking instead, which isn’t as good for position accuracy and has coverage limitations, but is fitted securely.
Retrofit: Positioning technology requiring an external power supply to provide constant power, or a GPS antenna that must be carefully positioned with line-of-sight to sky (e.g. under dashboard, not under the hood) is wholly unsuitable for mass adoption of secure road pricing systems. The cost of retrofitting millions of vehicles with expensive tracking hardware would be immense, and also incur liability risks.
It would cause uproar that governments are installing surveillance equipment into vehicles. Drivers will be most unhappy having their vehicles intrusively modified. By being so physically installed, it’s a ‘bigger thing’ than simply placing a small toll transponder on their windscreen and never worrying about it again.
Communication: GPS is point based tracking, requiring a lot of data to transmit routes. Dynamic road pricing requires high resolution of positions recorded and sent, unlike normal security asset monitoring. They need a solid internal power source. Neomatrix compresses routes so small they can be powered externally (no battery) via RFID like current toll transponders to send a full route history. Neomatrix compresses routes to a few dozen bytes, not several Kb (two orders of magnitude smaller).
For reference, similar new tracking technology Sigfox (Still antenna based) reports: An uplink message has up to 12-bytes payload & takes an average 2s to transmit to the network. This is sufficient to send 2x GPS points at 3m accuracy. However for comparison, Neomatrix can send a 30min full route in just 20 bytes.
Security: GPS systems are defeated easily, with lots of internet examples. It’s semi-common practice among some commercial drivers to disable tracking of vehicles or their containers to avoid monitoring. With the need for an ‘accessible’ antenna and power cable, such devices can be very defeated by severing wires, shielding or damaging antennas, or using (illegal) GPS jamming devices which are very cheap, and used regularly. This causes problems – even drone crashes.
At extremes, it’s easy to disable GPS satellites by rouge nations firing ASAT rockets to disrupt economies which may sound far fetched till we realise the fragile economy with Covid disrupting global supply lines so rapidly. GPS Spoofing can also affect wide areas, reported by the BBC as used in Russia.
While unlikely, the more we rely on GPS, it becomes a weak link target, which is specifically why the EU and others have felt the need to spend billions to develop their own systems, to address the issue of weak link reliance.
Cost: With millions of vehicles, GPS at scale is expensive, let alone significant installation costs and associated liability. GPS is cheap individually, but not compared to alternative positioning technology such as developed, which is far cheaper, simpler, and needs no antenna or installation.
Privacy: GPS resolves position to an error of maybe 3m, so it is high-resolution tracking. It is an obvious concern for drivers, who feel this is intrusive. The purpose of tracking is solely to detect and price use of main roads at busy and/or congested times, yet the government is tracking their movements everywhere, anytime, and accurately. It seems an extreme overreach.
The government may say ‘Yes we could track you in private areas, but we promise we won’t do this’ would be endless debate and concern in consumer’s minds that they are entering an Orwellian surveillance state. Drivers don’t want to be tracked when having affairs, spending their work time at the beach, or perhaps even speeding on quiet roads. Even when conducting business normally, drivers will be ‘disconcerted’ knowing they are tracked to such high resolution and all the time.
Alternative, new technology addresses privacy in novel ways, with technology that is ‘physically incapable’ of tracking movement, except on main roads, and known/defined congested peak hour times which is the purpose of the data capture. Drivers will accept privacy implications better if it’s limited to only the purpose designed, which is a core GDPR data privacy requirement too, limiting capture to only that needed..
As a thought experiment, the government could easily erect many toll booths or plate monitoring cameras across all main roads to implement large scale dynamic road pricing. However, erecting toll booths each km along main roads is cost prohibitive, and visually unsightly. Still, this would not invoke extreme privacy concerns that GPS would. Tracking ‘on’ the car or person is vastly more intrusive than a fixed toll booth tracking a car passing by. Users easily understand the reason for the toll booth tracking, but can’t justify individual car tracking constantly ‘in case’ it may drive along a toll road (or priced road) at some future time.
To address privacy well, monitoring systems should only track cars when actually driving busy main roads, but not to the high resolution of GPS, or to the global coverage area of GPS (everywhere).
Neomatrix positioning can only do main road ‘route tracking’, not point tracking (like GPS), so allows privacy to be addressed by default – as it is only able to monitor movement along main road routes.
Evidence on privacy concerns: A significant UK Government paper called London stalling. Reducing traffic congestion in London includes these comments from David Kurten AM, UKIP Group Lead on the Transport Committee: “The ultimate aspiration for London however is ‘big brother’ style total vehicle monitoring for the entire Greater London area. All vehicle movements would be monitored and charged by a government agency: probably TfL. This will destroy privacy & civil liberties for motorists in London.”
Comments above show a possibly incorrect understanding that vehicle monitoring must be ‘all vehicle movements’, meaning point based (GPS). Route based tracking is very different, and almost an extension on more toll roads, but without needing any toll gantries.
Privacy is addressed by a radical shift in how we monitor cars. Existing tracking systems are ‘Big Brother’ and cover ‘entire areas’, when this isn’t needed if a new system can only physically monitor main road driving. Route tracking is also technically unable to identify end points of journeys, as they are not ‘a route’, being another benefit for privacy over GPS which has no difficulty knowing exact end addresses.
GPS is a point tracking system and provides high accuracy, high resolution data on vehicle speed. This is frequently used by fleet managers and parents to take action against offenders.
For mass adoption of a road pricing scheme, the new ability of authorities to instantly detect all instances of speeding with accuracy is a worrying concern for big brother. Speeding isn't a factor in causing road congestion, and indeed while illegal, it could even be seen as helpful for reducing time spent on roads. Strategies to reduce congestion should not be confused with, or integrated with initiatives to address speeding.
With public sensitive to big brother privacy, 'mass surveillance' of citizens for purposes of reducing congestion - systems need to take care not to bleed into areas they were not originally designed for.
Because the Neomatrix method isn't a point-based tracking system, it simply can't be used for this in the same way as GPS. It can't reliably detect isolated instances of speeding. Additionally, it can't detect low level speeding in metro areas. The monitoring method simply can't 'see' these events clearly. Only prolonged speeding can be detected, exactly how point-to-point speeding cameras work. The public seem more comfortable with these schemes than being caught at a specific 'point', and want the ability to e.g. speed in order to overtake, but not necessarily have this detected as an offence.
The Neomatrix method can't detect speeding easily unless prolonged, but also, the NeoTag format smooths out speeding as part of its compression method. The data received from smart number plates by road authorities is thus even more broad, with individual and isolated speeding not seen clearly to begin, and also further smoothed if it is detected.
The end result is that unlike a point tracking system like GPS, Neomatrix positioning directly addresses privacy concerns on speeding detection. The system can be used to detect speeding but only in limited cases of extensive or prolonged speeding, and is far less likely to detect speeding in metro areas due to the smoothing methods. The public may be more accepting of this use case, but essentially, the system is designed to reduce traffic congestion, not to be used as a big brother surveillance tool, such as mass speeding detection. That it can't be used effectively for such surveillance should offer some comfort.
For additional privacy, the city can define a limited set of roads to price, rather than have all roads subject to monitoring, but many charged at £0 if local etc. This is to say the city can control where the system can technically track drivers, to provide transparency in them only being subject to monitoring on main (priced) roads, but not essentially all roads as is the case with GPS.
For additional privacy, a blockchain solution can be setup which can be reviewed by external parties. This can provide irrefutable proof that monitoring can only occur on designated main roads which are subject to a price for use (at relevant times). This is very different to GPS, as once signals are received, it’s hard to ‘prove’ they are not being misused. The NeoMatrix technology doesn’t receive GPS points as such, so has a technological dependency that can be proved doesn’t exist on the device, but which can only monitor where appropriate as the blockchain approach can prove this.
There are solutions which enable resolving the fee for the road use on-device. As such, the in-car device communicated the bill to be paid for road use, and never the tracking details. This scheme is not onerous or technically that difficult. It fully supports occasional updates from road authorities on pricing schemes, as well as methods to verify the bill owing claimed by the device/car. Of course the price, and any occasional network challenge issued to the car are also highly secure.
Video camera plate surveillance is an alternative to using toll transponders and toll booth gantries to detect vehicle toll tags. It is very expensive to install and maintain cameras, which means they can't cover all main roads effectively. It is also somewhat onerous for privacy concerns.
The recent London report on congestion as clearly stated that automated number plate recognition was dated, and that new technology to monitor vehicle use was needed.
For road pricing, vehicles must be monitored. Studies consistently advise that vehicles should be monitored per km and per time through main roads. Toll booths, automatic number plate recognition and congestion zones are regarded as 'Not fit for purpose' according to a recent and extensive UK government report on congestion in London. This report states that new monitoring technology for vehicles is needed to support road pricing effectively.
Retrofitting millions of existing vehicles is very expensive, and will have significant push back. Owners don't want government 'surveillance' hardware installed into their vehicles. Questions will be raised if this device records conversation for example. Issues of liability, who pays for the power, depreciating car value, damage, installer accreditation and so much more. In short it would be a PR nightmare to sell an installed tracking system to a city.
Road pricing should focus on small, cheap, removable devices that do not require installation or vehicle power. The public are happy to install a toll transponder on their windshield, or a secure smart number plate. Neomatrix technology can be used in the same way, with smart toll transponders that detect position without toll booth gantries.
Neomatrix has developed a novel and useful method of highly compressing route information. This compresses a series of GPS points to a NeoTag, which is just a few characters. Compression rates of 99% are possible over traditional methods, yet the few characters in a NeoTag can be decompressed to reveal the exact route, including timestamps.
Routes are not stored in a separate database like www.tinyurl.com, rather, they are stored within the tag itself. At present, testing of a large number of routes reveals this requires around 5 bits to encode 1km of travel, including full route (lossless) and time, which is significantly less than any prior art method such as Google Polyline or the Douglas–Peucker algorithm (lossy) etc.
Current methods remove redundant nodes from GPX files, or approximate paths with algorithms. Google uses an efficient Polyline method for it's online mapping applications but the NeoMatrix method shows a 70% better compression rate to this, for the specific use case of compressing road based movement.
Efficient data compression of routes may seem trivial, but is highly critical in the context of providing position visibility to IoT devices, which may have severely limited battery capacity, and storage space. Additionally, the incredible small compressed size of routes saves communication bandwidth, and also increases the reliability and communication 'windows' when sending data back to base - including sending via externally powered RFID (like toll transponders do)
There are many reasons efficient route compression is helpful, but the primary use case is for road pricing. Smart number plates or toll transponders may need to communicate a historical route taken, so use of congested roads at peak times can be sent to the toll authorities for billing.
In this context, there are very tiny batteries, as these devices can't be hard wired to vehicles for security, cost and other reasons. NeoTags are so incredibly tiny, which is helpful for a small battery powered monitoring device to send just a few characters to communicate a lot of driving. Also, connection windows are also tiny, as vehicles speed under toll booth stations - and can't be paused to complete a longer transmission.
By comparison, a legacy GPS tracking system can’t achieve anywhere close to this. Point based tracking would need to be resolved or snapped to roads prior to sending, but this is an onerous computation process that has a high power draw. As such, snapping to roads is done server side remotely, but then this requires communicating a long series of GPS points which is a problem. The NeoMatrix method doesn’t work in points, so has no need to preprocess ‘snap to road’ before sending compressed route data.
The NeoTag method is complex, but is perfectly designed for road pricing schemes which is what it was specifically developed for.
While technical aspects of road pricing are a significant barrier to adoption in monitoring positions, consumer acceptance is also a big hurdle to address.
Privacy aspects are critical. GPS and any point tracking system is highly invasive for privacy. Consumers don't want to be tracked 24/7 or to high resolution. They don't want to be detected occasionally speeding, and they don't want their final destinations tracked.
Neomatrix technology addresses all these concerns. By only being able to effectively track main road use, being unable to track final stop destinations accurately, and being unable to detect speeding unless sustained … Neomatrix gives far more privacy confidence.
How a scheme is described is critical. Language is key. Sell the benefits, not features.
Negative terms: Congestion tax, road tax, road pricing, road tolls..
Positive terms: Road pass. Peak hour passes. Fast Pass, Fast Lane pass. Speed passes.
We are 'selling' the opportunity to travel at peak hour faster than now. Road pricing reduces congestion - so consumers are buying time, and thus traversing the network faster.
The word tax implies a sunk cost. We may not see benefits of this directly, and not as an individual. A road fast pass however, seems to offer a direct and individual benefit to the driver. They are paying to get somewhere faster. They get something for their money.
Minnesota USA has an express lane scheme: www.dot.state.mn.us/mnpass/howmnpassworks.html It is marketed in a positive manner as an ‘express lane’ or ‘fast pass’. In this case however, it is only one designated lane that is run with road pricing. In non-peak times this lane is open to general use for no charge. This is a little confusing for consumers, but also quite inefficient. Far better is to have all lanes designated as subject to road pricing, even if the price is zero for off peak travel.
Done well, road pricing can easily be a win-win, reducing costs overall, for both drivers and also government. The scheme seeks to primarily modify driving behaviour slightly, not introduce additional fees. Even drivers who spend a lot of time in peak hour can save if the cost of road use equates to reduced congestion, and thus more time earning. Others will not pay anymore provided they can adjust departure times a bit perhaps.
The ability of road pricing to provide a feedback loop is excellent, but consumer comfort regards privacy is critical. Thus, having a scheme which is technically only capable of tracking routes (not everywhere), and can be set to only track billable routes, consumers are far more likely to accept it more happily.
February 2021