Tag: Future

Location based road tolls

 

This is a pre proposal for a location based road toll system. It’s modified from one that I wrote in Finnish for a local discussion on the subject. It’s “pre” in a that it doesn’t even try to be complete, rather privacy and reliability against cheating are considered. Privacy from both the perspective of real time location as well as from the perspective of a location archive. Location is defined with the extra attribute that it is collected in a distributed manner, the car (or users device) collects it based for example on GPS or other satellite locationing system, or for example by a LIDAR.

See also Jakke’s take on the subject.

That governments are inventive in creating new things to tax is taken as a given. It is not my intention to consider what would be the correct toll level of the toll at any time, location, vehicle type, weather etc. The reason to have a toll system that covers all public roads and has a road use cost that changes depending on conditions is efficiency. Limited resources are available for example to road building and maintenance. By setting the price of road use higher at rush hour it is possible to control the number of people that decide to use that part of the road at that time. In short the system creates a market for a good. The reason to have a location based system is that it has very fine granularity, for example it is able to set the price of using a road near a school higher and thus able to give an incentive to select an alternate route.

Premises:

a) The system doesn’t need to be perfect. Someone might be able to cheat and not get caught. This is true with most rules, there is no need to catch every shoplifter or tax evader to keep the correspondent problem at bay. To give some context: in March 2012 a the newspaper Helsingin Sanomat reported that in  Finland the police is able to solve about 42 % of all property, violence and sex related crimes.

b) The ability of (local) governments to manage software and technology projects is limited (this may be true universally, but I know its true in Finland). Thus it would be beneficial to use private resources to develop the system.

c) monitoring of compliance should be done in the traditional way, i.e. spot checks. This may be automated, but the number of checks should be limited.

d) For those that have difficulties to use new technology or oppose it on principle an alternate way of paying is offered. For example a ticket that is valid for a day or a month. For a large majority this should be more expensive that using the location based system. When buying such a ticket some routes may be restricted. These limitations are needed to make most people use the location based system. Otherwise the market would not function properly.

e) A fairly reliable wireless network and cheap devices to use it are available.

Technical details are not considered in detail as they are considered to be fairly easily solvable. Some changes to laws may be required depending on the jurisdiction and at least in Finland there will be an outcry if a private company collects payments resembling taxes.

Figure 1 shows a simplified flow of system operation. Vertical dashed lines partition the operations done by the three different stake holders. The boxes marked “SW” show that the driver gets the software from the company, for example through an app store.

The third party (i.e. “company”) is brought to this picture mainly because I believe that it can develop and operate the system more efficiently. Secondly it is much easier to take ones business elsewhere than to do the same with a government. Thirdly losing consumer trust is a bigger thing to a company than losing trust of citizens is for a government.

At “Start” either the car or the driver starts the system. Several different kinds of automatic ways of doing this are possible for example Bluetooth connection with car can trigger the system. The information that recording has started is transmitted to the company. Location may be sent with the information but this might not be mandatory as only information that the system is recording could be enough at this point. At set intervals (location or time) the location is recorded. It is possible to encrypt and transmit this information at longer intervals, this information could be used by the company to check that the final trip info (sent later) is correct. At “End” the location information is sent to the company.

The tolls are defined by the (local)government or the owner of the road. The “Price” may be dependent on time, weather, location, congestion or other parameters. It is possible to have continuously varying prices, but at least human (in contrast to robot) drivers may prefer to have the prices before the journey.

A HASH is calculated to make it possible to check that the information has not changed after it leaves the drivers device. “Encrypt and save” the data to the users device for further use. Unencrypted information is transmitted to the company so that it can check that the trip appears to be a valid one.

If it could be assumed that the software or hardware has not been tampered with, the information could be transmitted to the company in encrypted form and the possibility of a privacy violation at the orange “Check” could be avoided. I don’t believe that this is possible without either using expensive custom hardware or severely restricting the users ability to use the device. Another way could be to transmit only encrypted information but trust the checking mechanisms introduced below. This however could lead to more frequent point checks which would be an other kind of privacy violation.

When the trip information is received by the company, it is checked for consistency. For example are the location points sensible or are some located in the middle of the Pacific ocean or are some points missing? If there are some problems the user is notified and the information about the problems is archived. For those who have a lot of problems actions can be suggested, to update devices etc. If everything is fine, the information is encrypted and the unencrypted data is destroyed. Encryption is done by using the public key of the driver. This means that no accessible database of driver location is formed. The data can only be accessed if the driver gives his private key.

Billing is done as indicated in the figure, at the same time the encrypted info is copied to the governments archive.

For those who value their privacy more than others it is possible to sell prepaid codes that are connected to an account on a company database. The driver could use for example a Tor network to update the location information to the company database. This way the company would at no point know for certain who is travelling. Of course if the trips start at home or work it would be fairly easy to match this information to at least a group of people.

Figure1_roadtolls
Figure 1. Simplified system operation.

In addition to tampering with the soft- or hardware it is possible to trick the system by simply not initiating a journey when necessary. To be able to rely on something else than trustworthiness of the users Figure 2 shows a procedure to check that the user is doing his end of the deal.

The inspector asks the driver to stop. The driver then gives the inspector a code from his recording device. This code is given to the company which returns with information on the device’s status, either a trip is ongoing or not. Depending on how strict the privacy is the company could also give information on how long the trip has been going on or what has been the distance travelled. It could for example tell if the distance is over one kilo meter. By following the driver for a short while this could rule out cases where the trip is started only after the request to stop has been given.

It is likely possible to automate the checking procedure for example by reading the licence plate using machine vision and checking from the company if a car with these plates is on a trip.

Checking of driver
Figure 1. Checking of driver

If the company income is tied to the amount the driver pays incentive for the company to cheat is diminished. This is true especially if the invoice makes a trip through the government bureaucracy as this makes it possible to check that the sums are statistically in the ballpark. Figure 3 shows how the company can be inspected without violating the privacy of the driver. The idea is to pay (random) drivers for revealing their whereabouts. When voluntariness and an auction is used, no one needs to yield unless the compensation is adequate.

The encrypted information saved in the users device is compared to those the company has given the government. If this is done by the driver the government never knows where the driver has been. Because these inspections would hit a company fairly often it can be assumed that enough honest drivers are reached to be able to determine if a company is following the rules.

As the road prices need to be public it is possible for the driver to make certain, perhaps by using free software, that the company is charging the correct price.

Figure 3. Checking the company.
Figure 3. Checking the company.

Future with driverless cars 6: Zoning, Traffic volume and Externalities

 

This is part six of a series on changes that driverless cars may bring. I expose some of the ideas I have, mostly quite practical things. We moved project troglodyte to it’s own website, so the more patent centric recap of the Google driverless car patents can be found from there.

Below I assume that the problem has been solved completely. Driverless cars can access any part of the road network, function even when there are people darting around and can handle any weather including lots of snow and very slippery conditions. Accident levels are same or lower than currently and people are not scared to use autonomous cars.

See also: Rental vs. taxi, Mass transportation, Pirvate cars, Cargo, Parking and driving empty, Zoning Traffic volume and Externalities

In smaller scale considerable flexibility will be created when parking can be some distance away from destinations. In densely populated areas parking facilities can be concentrated to avoid parking on the streets and the most valuable lots can be used more efficiently. In areas where snow cover makes street maintenance difficult cars can be either automatically moved to allow for removal of the snow or parking on the streets can be prohibited.

Removing parking from some areas makes it possible to create very densely built environments where walking is an attractive alternative for moving from one place to another. Denser areas also strongly support public transportation as the total trip time will be low due to short walking distances.

On a slightly larger scale putting commercial, industrial and residential districts close together would enable efficient sharing of vehicles as the same vehicle could transport several persons during one rush hour because the time driven empty would be short. In case of a city center/suburbia structure, serial sharing would also be possible if the rush hour peak is flat enough, but that would increase the total distance driven due to the long trip back to suburbia to pick up the next person.

In areas where the rush hour traffic is mostly one way and the road has several lanes to both directions using more lanes to the direction that has more traffic would increase capacity with relatively minimal investment to infrastructure. This is already possible with human drivers and even in use, though quite rarely. Using automated cars would make it much more flexible. Depending on how this is implemented it might be  necessary to forbid human driven cars from at least some lanes.

Automated cars can take kids to school. This would make it easier for people to live in very sparsely populated areas as parents would not need to drive their kids to school. Several kids could naturally use the same car when they know each other. As a downside parents might put children to schools that are quite far from their home. While this might have a leveling effect on house prices it would make the kids spend a lot of time travelling. The effect on housing prices is due to the fact that parents are willing to move to areas where good schools are available. This is apparently an important factor creating price divergence between areas.

Overall traffic volume would likely increase as driving would be easier, autonomous vehicles would be able to drive by themselves and new groups of people could use private cars. On the other hand relative price competitiveness of public transportation especially busses would likely increase.

People seem to be fairly bad at estimating how expensive driving is, this is particularly true before they make decisions that affect their driving needs for years to come. So costs of car maintenance, operation and time spent in traffic jams have a delayed effect on behavior. Because of the holding pattern parking problem an environment with many automated cars needs some sort of congestion charge system that is dependent on the distance driven and on the current need of that part of the road network. If electric cars become popular and the price of solar panels keeps plummeting, fuel cost might be close to zero in some cases. This would amplify the unnecessary use of road problem if no separate charge for road use exists.  Knowing that every kilometer is charged might have an effect on how willing people are to use a private car and lead to a lowering of the traffic volume.

The most common externalities of traffic that are affected by automating driving are land use for roads, adding to climate change and noise.

If the traffic volume does not increase significantly then the growth in capacity per road area brought by automation could lower the need for land under roads. It could also be possible to use smaller intersections especially on faster roads as traffic on adjacent lanes could have larger speed differences.

A major contributor for traffic land use change could come from the relocation of parking space to less valuable areas. Increases in popularity of public transportation will likely lead to higher throughput for the road network.

Noise is dependent on the number of cars but also strongly on their speed. Automated cars will follow speed limits so lower speeds can be used in areas where the harm from noise is large. Automated traffic can also lead to less accelerations which produce more noise than steady speed. Automated cars travelling empty can drive very slowly to reduce energy consumption.

Acknowledgment:  Thanks to Laston Kirkland for thoughtful evaluation of these ideas.

Future with driverless cars 5: Parking and driving empty

 

This is part five of a series on changes that driverless cars may bring. I expose some of the ideas I have, mostly quite practical things. We moved project troglodyte to it’s own website, so the more patent centric recap of the Google driverless car patents can be found from there.

Below I assume that the problem has been solved completely. Driverless cars can access any part of the road network, function even when there are people darting around and can handle any weather including lots of snow and very slippery conditions. Accident levels are same or lower than currently and people are not scared to use autonomous cars.

See also: Rental vs. taxi, Mass transportation, Pirvate cars, Cargo, Parking and driving empty, Zoning Traffic volume and Externalities

Driverless cars can park themselves. The driver can be left as close to the destination as she wishes while the car will find a parking spot optimizing the cost and the time it takes for it to be available for the next occupant. Parameters like the usual time the car is next needed can be taken into account when the decision is made.

When there is no need for humans to normally enter the car in the garage, cars can be closer to each other. Cars that have been scheduled for next assignment can move inside the garage as the situation changes so that they can leave as close to optimum time as possible. To enable efficient operation there will need to be a possibility for the car to take instructions from the garage computer or equivalent.

Some lowering of cost may be gained from designing the garages for vehicles only, they will need less ventilation and emergency exits. When no one is around more aggressive fire suppression methods can be either remotely used or automated. Essentially a garage will be a big machine where other machines enter and leave.

Paying for the stay must be automated as well. This may produce some conundrums as the car is in effect buying stuff by itself, but similar safeguards as used for example in Facebook marketing may be used, i.e. a maximum per minute and a maximum for a stay. As the car can move to a next garage parking space may be auctioned continually and if it becomes too expensive the car will leave and find a cheaper place to stay.

Building tunnels is cheaper if human safety does not need to be considered. Automated cars may lead to building of “cars only” infrastructure, where a vehicle carrying passengers may not enter. Empty vehicles could leave a point of interest such as a stadium through a small tunnel that is not equipped with emergency exits etc. This would allow the cramming of more capacity to smaller space in densely built areas.

Acknowledgment:  Thanks to Laston Kirkland for thoughtful evaluation of these ideas.

Future with driverless cars 4: Cargo

 

This is part four of a series on changes that driverless cars may bring. I expose some of the ideas I have, mostly quite practical things. We moved project troglodyte to it’s own website, so the more patent centric recap of the Google driverless car patents can be found from there.

Below I assume that the problem has been solved completely. Driverless cars can access any part of the road network, function even when there are people darting around and can handle any weather including lots of snow and very slippery conditions. Accident levels are same or lower than currently and people are not scared to use autonomous cars.

See also: Rental vs. taxi, Mass transportation, Pirvate cars, Cargo, Parking and driving empty, Zoning Traffic volume and Externalities

Transportation of containers between terminals is an obvious place for automated vehicles. Both the loading and unloading can be automated and the vehicles can create a rolling conveyor belt. Specialized vehicles exactly the length of the standard shipping container will likely be built, these vehicles can form trains on faster roads and thus lower their energy consumption considerably.

For cargo travelling shorter distances and requiring manual unloading upon delivery drivers might still be necessary, although they might not actually drive the vehicle. But even here a driver may not be needed. With some development it might for example be possible to unload a truck using its own truck mounted crane by remote control. The crane operator sits in an office and directs the crane using cameras and a communications link. There might still be problems like how to judge how good the support is for the vehicle when the crane needs to move a heavy load far away from the truck.

Remote operation might also be usable when the environment at the delivery or loading point is not easily standardized. Examples might be a quarry or a landfill. The vehicle would be autonomous on the road network, but remotely controlled when when the environment is difficult. This would make it possible for one driver to control several vehicles. The same operator could also control the loading equipment for example a wheel loader.

For light cargo: delivery of pizzas, groceries, stuff bought from the net etc. automated vehicles can be a real revolution. A suitably equipped vehicle may quite easily open a container when a code is given and the customer can take what was delivered. Significantly, driverless cars may make it cheaper to order a carton of eggs from the store than to drive there and buy them. Products for several customers can be delivered in the same run and the cost divided between the receivers whereas when driving to the store you pay for the whole trip. Currently in many cases the cost of the delivery vehicle driver tips the balance the other way.

It may be a bit futuristic to think that specialized pizza vehicles able to make the product would drive around the suburbs, but it is by no means impossible. Making of a pizza is not that difficult and delivering it fresh instead of 30 minutes old may make all the difference. The pizza would be ordered through the net and the pizza van would start making it so that it is right out of the oven when it arrives at the delivery location.

Acknowledgment:  Thanks to Laston Kirkland for thoughtful evaluation of these ideas.

Future with driverless cars 3: Private cars

 

This is part three of a series on changes that driverless cars may bring. I expose some of the ideas I have, mostly quite practical things. We moved project troglodyte to it’s own website, so the more patent centric recap of the Google driverless car patents can be found from there.

Below I assume that the problem has been solved completely. Driverless cars can access any part of the road network, function even when there are people darting around and can handle any weather including lots of snow and very slippery conditions. Accident levels are same or lower than currently and people are not scared to use autonomous cars.

See also: Rental vs. taxi, Mass transportation, Pirvate cars, Cargo, Parking and driving empty, Zoning Traffic volume and Externalities

While there is a clear reduction on operating cost for taxis and public transportation, private cars will only get an initial premium to purchasing price. Economic benefits will be less direct, but as mentioned in another post renting the car out when it is not needed is clearly a possibility. It is difficult to tell if  this will drive taxi and rental car operators out of business or make more people start using public transportation and rent when traveling outside the coverage area of scheduled public transportation. The optimum solution will likely depend on how population is distributed in a certain area.

Not needing to drive the car is a benefit for many, but not all. According to wikipedia depending on the severity of motion 33 to 66 % of people are susceptible to motion sickness. This limits the number of people who can read or work during a car trip and thus also limits the benefit from a driverless car. Some people also enjoy driving a vehicle and wouldn’t want to give it up.

Not everybody can drive though and these individuals have most to gain from this development, in addition to public transportation they can have the option of owning a personal car. Beside some adults either unable or unwilling to drive, children and elderly are a large group that cannot or should not drive. As an example children under the age of 15 and elderly over the age of 75 make up about 25 % of the population in Finland*.

Adults under the influence of alcohol, medication or recreational substances is a group that will greatly benefit from automated cars for their transportation. There might be adverse health effects as being able to drive might be a sufficient reason to stop drinking after the weekend and this reason will be less powerful when there is no need to worry about driving.

Private cars are needed a couple of times during the day to go to work, get back home etc. These are also times when there is maximum need for transportation, so scaling the taxi fleet to meet that demand is likely not economical, even when the cost of the driver is removed from the equation. Many of the taxis would be idle for the rest of the time and not producing. Although this is probably the situation now for many taxis and certainly is for private cars.

Even if the rush hour peak in transportation need would be flattened this might not produce efficient use of the available cars in a case where private cars would be rented when not used by their owners. If housing and places of employment are at separate areas it takes some time for the vehicles to travel back to pick up the next passenger. Thus it looks like densely built areas with mixing of commercial and residential districts may benefit from a change to driverless vehicles. That however requires that the peak is wide enough to allow a vehicle to transport several consecutive passengers during one rush hour.

While vehicle size for public transportation is driven partly by the cost of the driver there is no such link for private cars. People usually want a car large enough that they can pack the whole family, including the dog and a canoe in to it. One reason for large vehicle size is safety, a large car with big mass experiences statistically smaller accelerations in accidents and can be safer than a smaller one. A big car likely also feels safer even if it might have inferior technology and actually be less safe. If automated cars live up to their promise and accident rates become lower there is the possibility that consumers may accept smaller cars.

Energy efficiency and operating cost are currently factors that drive private car size towards smaller vehicles. When the car can also drive by itself it is possible that the yearly distance travelled by the car will increase, this would also put emphasis on the  operating cost and make smaller cars more desirable.

It is also possible that a super mini, i.e. just one person, class emerges to enable sending the car on an errand by itself. The car could for example collect food from several places and bring it back home. But to be viable a car this small would likely also require some more traditional use to be economical.

Parking is a problem in many areas. It is infact an important reason for the use of public transportation. Here driverless cars will have a considerable effect. Currently parking space needs to be within a walkable distance from the place where the driver is heading. With driverless cars it is possible to drive to the closest point accessible with the car and the car can then find a parking space on its own.

While this will eliminate driving around to find a parking space, it can increase traffic close to points of interest and increase the overall distance driven as the cars may travel fairly far when optimizing between price of driving and the price of the parking space. This may lead to situations where for a short visit the vehicle is left to drive around so that it is almost immediately available for the owner. This in turn creates a need to put a price for the use of the road network, otherwise it might be cheaper to use a holding pattern than to park.

It might still be the case that every now and then an unseen situation is found which the car can not handle on its own. Remote operation may be used in these cases if the vehicle either does not have controls or the occupants are unqualified to operate it. For control only in rare circumstances basic controls may be enough, similar to game consoles or just an application using a touch screen. This off course requires upgrades in communications infrastructure as the road network currently covers some areas that have bad or no affordable ways to transmit for example video streams.

Acknowledgment:  Thanks to Laston Kirkland for thoughtful evaluation of these ideas.

Translate »