You can improve your parking time, reduce fuel consumption and help the community to efficiently plan parking lots!
Problems & solutions
When I'm searching for a free spot is hard to find the closest because of other obstacles that are in the way or to verify if my car really fits. So I'm thinking Augmented Reality should be the most intuitive: just open an app on your smartphone ( or OpenXC device from your vehicle ) and scan through your camera, immersed in the environment, where you can find the next free spot!
The idea here is to intelligently track each parking place. For this I'm assigning a specific address & GPS coordinates to each place, so that I can find the exact position.
More, I need that the system to easily measure each place free region - so I'm using a ultrasonic radar setup that can render a minimal polygon for each parking place ( you don't have to worry if next driver has badly parked his car, or if some small non-metallic obstacle is in the way, the system will report just the free region, and if your car fits, then go ahead! )
To be cheap to implement and maintain, and also to have minimal data transport, I've organised the whole infrastructure like so:
- Each geometric partition of parking lots ( a row, an arc etc. ) has slave-addressed tracking devices consisting of 360 degrees ultrasonic radar, each of them chained on a wired RS485 connection
- Each such partition has an WiFi master node which talks with its slaves and then publishes status in cloud
- The best thing here is: MQTT protocol; each master from 2 has a MQTT client which publishes data to a broker
- The MQTT broker then sends current status to each subscribers ( like smartphones, OpenXC devices etc. )
- On the server each parking lot has its boundary acting as a geo-fence and based on position of the vehicle that searches a place and the topography / occupancy it uses a simple Pathfinding algorithm to resolve next optimal free place.
Long story short
You can find a free space to park your car much faster, more intuitive by using AR, and the whole system can be implemented / maintained cheaper than the current alternatives ( for eg. Streetline ).
This project was created in 2-3 weeks ( from concept, to POC and simulations, to actual working prototypes and apps ) for the Ford's challenge Parking Lot 2.0: http://parking.challengepost.com and it missed the deadline for submission by a few seconds so now it became a portfolio presentation. :)
For extra info about software, hardware, algorithm pls ask...