How can the Internet of Things be used to improve drivers’ parking experience? Is it possible to make it easier to spot a free parking space in a garage? The AUTOPILOT webinar on 31 May explored these questions and many more. The second in the AUTOPILOT webinar series, it attracted over 70 listeners. The webinar was hosted by TNO and focused on the work being done at the AUTOPILOT test site in Brainport, Eindhoven region.
Sven Jansen (TNO) mentioned that several AUTOPILOT project partners are involved in the work on parking applications that is now happening in Brainport – DLR, Vicomtech, NEVS, TASS International and Gemeente Helmond. The work is led and coordinated by TNO. Mr Jansen explained that, along with automated valet parking (AVP), partners in Brainport are working also on technologies that enable car sharing, platooning, highway driving, and vehicle relocation. In AVP (development supervised by DLR), the vehicle is enabled through IoT to drive unmanned to a parking spot and to return to the driver on command. This service brings a range of benefits linked to efficient use of space and time, and even optimisation of logistics processes.
Emi Mathews (TASS International) then took the floor to explain the technical features of AVP. First and foremost, as Mr Mathews outlined, there is the IoT Platform, which allows IoT devices to communicate with each other using Internet protocols. Then, the vehicle is equipped with an in-vehicle IoT platform, which provides a range of services, including parking manoeuver service and motion planning service. The IoT Gateway in the in-vehicle platform interacts with the overarching IoT Platform to receive data from external sources. Mr Mathews pointed out that the AVP app itself is yet another service. The app can start or stop the AVP service, send commands to the vehicle to retrieve it or send it to the parking lot.
Next, Jorge Garcia (Vicomtech) spoke about the use of video surveillance, which in AVP is used to identify available parking spots. The technology is based on a deep learning model, using a computational neural network and adding a classification step on top. An application has been developed to send real-time information about the parking spot status to the IoT platform. The information then is available to the vehicle and it can park in the available parking spot. The main challenge here is that available parking spots may be occluded by some obstacle or another vehicle in the parking space. When this happens, the application is unable to know the status of the parking space. The application functions better when it has an overhead view that cannot be occluded.
Robert Kaul (DLR) presented the drone (micro air vehicle – MAV) application use in AVP. Here, the MAV itself becomes an IoT device, which is able to collect data by navigating autonomously when there is no GPS signal available. When an automated vehicle arrives at the drop-off zone, the driver gets out of the vehicle and, through the app, sends a request to the IoT Platform to start the MAV. The MAV hovers across the parking area, detecting empty parking spaces and occupied ones, and sends information to the app. The driver can then confirm the selected parking spot, where the vehicle proceeds to park itself. The MAV flies back to its base station and lands. The same happens vice-versa when the vehicle is being retrieved from the parking area.
Finally, Mr Jansen laid out the Brainport test site’s plans for the future. Two plug test events have already been completed, where the integration of technologies has been validated and the baseline functionality for the first testing phase has been defined. Now Brainport will focus on expanding the service concept of individual use cases and in 2019 move to the integration of the different services, looking at how and which services can be grouped together. The services will be demonstrated at the ITS European Congress next year, which will be held from 3 to 6 June 2019 in Brainport.
For more information, including the webinar presentation and recording, visit the Webinars page.