Image (870x540)

In order to evaluate and comment ideas you need to be part of the community.
Join community now.

additional attachment(s):

extension siemens_parking_problem.pdf

A decentralized solution to the parking space problem

Short description (max. 250 characters)
Much fuel is wasted by driving around while looking for a free parking space.
I propose a new way to guide drivers to free spaces without need for a central authority or fixed sensors to check if a space is occupied or not.

Key benefits (max. 250 characters)
- less waste of fuel while looking for free space
- decentralized approach that is hard to interrupt
- no initial costs for sensor distribution
- no running expenses for sensor maintenance
- open system that enables competition

Detailed description (max. 500 words)
Finding a free parking space in a crowded city can be a difficult and time consuming task. In larger citys up to 40% of the traffic is caused by people crusing around searching for a space. In Schwabing, the part of munich where I live, this sums up to about 40.000km every day(!), wasting gasoline worth 3.5 million Euros each year.
There are already several cities where free spaces are registered by sensors in the parking meters and then sold to drivers at demand dependent prizes.

In my opinion those approaches have several drawbacks:

- The spaces are detected by sensors with low range so that many of them have to be distributed all over the city. Since every sensor needs to communicate, some kind of telemetry has to be installed for every sensor. Due to the needed maintenance there's not only a large initial investment but also running costs that have to be included into the prize of the parking spaces.

- There is one central authority that decides on the assignment of the spaces and the prize. There is no competition in the market that could force the development of better assignment strategies or prizing schemes.

To find empty spaces, I propose to use the ultrasound sensors of current driving assistance systems together with the GPS and the data connection of the driver's navigation system or smartphone. Modern cars are more and more equipped with assistance systems that detect possible parking spaces with side facing ultrasound sensors and even pull in automatically. Since those sensors are very cheap I expect those systems to be a standard feature within the next years. If the driver is not searching for a free space this information could be published to other drivers via ad-hoc networks or stored in the cloud.

If a driver is searching for a space, he can download the information from the server and compute an optimal route to find a free space. Since the space is not reserved in this system, he has to take the possibility into account that his space is not free anymore if he gets there.
This probability depends on the position of the free space (does it lie in a small hidden street with little traffic or next to the main street) and the time to get there. This time dependent probability can be estimated using recorded data and machine learning algorithms. Graph algorithms allow to find a route that minimizes the time to find a free space. (Please check additional material for details)

To motivate drivers to announce free parking spaces, a reward system should be established. If a driver uses a space, he pays a small amount to the driver who found it. This could be implemented by using a social micro-payment service like Flattr.

All current projects can be implemented on top of this approach but without an own expensive sensor network. This system could even be started by a small group without financial backup or support from the city.

automobile, energy saving, mobility, traffic



Status: in discussion

111 times viewed

2 Comment(s)

Number: 000120

submitted: 18.07.2012