The lightning-scooter is an e-scooter whose electric drive can be booked for a certain period of time. For the payment Bitcoin Lightning is used. You choose on the display how long you want to drive, get a qr-code generated which you scan with a mobile phone lightning APP and pay. Then the power is switched on for the selected period and you can use the scooter. The special thing about this system is that it is mobile, i.e. the communication is completely realized via the mobile network.
On the 35C3 I offered the lightning-scooter for rent, it was a great success.
Here is a small video of the booking process:
In addition to my previous project, the lightning-bike, the lightning-scooter has an integration with streamr, an open-source platform for the worldwide exchange of real-time data. Now every time you make a booking, data about the lightning-scooter is transmitted to the streamr network, including date, location, battery charge status and much more.
It is thus possible to document the use of the lightning-scooter and to receive information about the scooter online via the streamr platform at any time.
The following information is transferred to streamr during every new rental process, in the next picture you can see the content of the stream:
In a test run I rented the scooter 4 times and then drove 1 minute, each time the status information was sent to streamr. From this data I created a canvas with the streamr editor which collects the data from the stream. It is now possible to visualize the GPS data in the canvas via the map module, so that the locations of the rent can be displayed. Since also date and time can be captured now e.g. motion profiles can be generated.
After the boot process, the client automatically starts the program on whose start screen you can currently choose between three different usage times, 1, 3 and 5 minutes. The costs per minute are 250 satoshi that is approx. 0.01€.
The customer now selects the desired rental time and the program first determines the exact position of the scooter via GPS. Then the current status data of the scooter is read out via Bluetooth. An internet connection is then established and the collected information is transmitted to the streamr platform.
In the second step, an invoice is generated for the requested amount, which is transmitted to the lightning node via lightning-charge. The client gets the payment information back from the lightning node and generates a qr-code which is shown to the customer on the display.
The customer now has 60 seconds to scan the qr-code with his lightning APP in his mobile phone and then pay the invoice. During this time the client tests the lightning node to see if the invoice is marked as paid.
If the payment is not confirmed within 60 seconds, the data is discarded and the program returns to the start screen. Here you have the possibility to repeat the process.
If the payment has worked, the Internet connection will be disconnected and the system will turn on the power for the selected time. The scooter is now ready to go! After the end of the paid time the power supply is interrupted and the rent is finished - of course it rolls on, but only with muscle power. The program returns to the start screen and is ready for a new rental process.
The heart of the system is a Raspberry Pi 3 A+, which is supplemented by a GSM/GPRS/GNSS HAT. This component has a GSM and a GPS module and is therefore responsible for the positioning as well as the Internet connection.
The relays for controlling the power supply are controlled by the GPIOs of the Raspberrypi Pi. The Bluetooth module reads the information from the scooter. Here I have adapted the program code from this repository ReadM365.
An e-paper display is used as monitor, which also provides 4 switches for the menu control. These are also queried via the GPIOs of the Raspberry Pi. The e-paper display has the advantage that it consumes almost no power in display mode but only when the image content changes. It has a high contrast and is easy to read even on sunny days. The picture construction is relatively tough with approx. 6 sec., but only 2 steps/pictures are needed to complete the payment process.
For the power supply of the Raspberry Pi it was necessary to install a DC/DC converter which provides the required 5V from the 42V of the scooter battery. In the lower picture you can see it, so that the converter does not discharge the battery, a switch is installed which separates it from the current. In this picture you can also see the two blue relays which are responsible for switching the main power.
Here you can see the wiring going from the e-paper display to the connectors on the Raspbery Pi. The connector on the right side of the upper shell leads the contacts of the relays and the current of the Paspberry Pi, which all come from the bottom of the scooter.
On the server side there is a c-lightning based lightning node. For the control of the lightning node the API lightning-charge is used because it's very easy to implement the payment processes.
The current raspbian buster lite distribution is installed on both Raspberry Pi systems. The program code to control the lightning-scooter is written in python.