Equipments we had at our disposal:
- 5 XBee S1 : chips that possess an antenna to send or receive datas
- 4 Arduinos Fio : chips that are programmable and can be connected to the Xbees
- 1 XBee USB adapter : to "talk" to an XBee with a computer
- 1 USB UART FTDI : to program the Arduinos
- 5 3.7V Lipo batteries
- 2 Arduino UNO
- 2 XBee shields : to connect the XBees and the Arduinos UNO (the link between the Arduino and the XBee didn't work but we could use them as USB adapter)
- A small remote-controlled car
- Several electronic components (resistors, capacitances, leds, a speaker, wires, one electronic board
The main goal of this project is to use Arduinos and XBees as a location system.
The first aim of was to use triangulation to be able to locate one chip using three others. The idea was to get the strength of the signal received by the three fixed chips and convert it into distance. We already saw in Belgium that the RSSI (received signal strength indicator) wasn't accurate, but we thought it was mainly due to the interferences and the reflexions because we live in the city.
But, when we took the measurements to calibrate our chips and find a relation between the RSSI and the distance, we were pretty surprized : even on a wide open area with very few interferences like the Martian desert, the RSSI kept a significant part of randomness. We were able to find a logarithmic function for distance -> RSSI that fits the measurements we got, but the errors between the estimation and the reality were pretty huge : for small distances, the RSSI changes a lot even if both chips don't move, and for greater distances, the small error in RSSI becomes exponential for the distance and we could have an error of 20 m! We knew it wasn't going to be very accurate, but this was way worse than what we thought.
We knew that this problem would jeopardize our project. While trying to find a solution, we tried to transmit datas between the XBees. It worked pretty well, it was faster than using an USB stick. It even worked from the observatory, which means that the astronomer could send information directly during their observations.
Among the several electronic component we brought, there are some LEDs and a little speaker. We first made a danger alarm : if the alarm chip (a chip connected to a speaker and a LED) gets a signal from another chip, it starts emmiting an alarm sound and the light turns on. The alarm chip starts to get a signal when it is close to the other chip, which would be about 100 m on a wild field or 50 m if there are rocks or hills between them. The danger beacon would be placed next to dangerous spots like cliffs and the alarm chip are supposed to warn the astronauts who are riding an ATV so that they can pay more attention or simply turn around. We tried it on an EVA with the ATVs to see if you can be warned on time and it worked pretty well.
Then we build what we called a "mobile traffic signal" : with three LEDs (green, yellow, red) connected to one chip, we could command them wirelessly. We thought of several applications but none of them was really helpfull. We improved this by switching on the different LEDs depending on the RSSI. Then, we added the speaker wich plays different tones depending on the RSSI also. So, even if the signal strength isn't accurate, it doesn't matter for this use: when you get closer to an emmiting chip, the leds change (from none to green to yellow to red) and the speaker plays a higher tone. It could have two major uses : we can put one in an interesting/dangerous spot and you know when you are close to it. The other use is that each crew member should carry an emmiting chip. If one of them is in danger and can't tell the other members of the crew where he is (becaused he passed out for instance), the other members can find him very easily. We tried it also and it worked well, but sometimes it is really hard to hear the speaker through the EVA helmet.We called this chip (the one with the three leds and the speaker) the "location chip". We improved it and the "beep beep" played by the speaker becomes shorter when you get closer. On our last EVA, one of us hide one chip, and the other one had to find it using the location chip. He succeed every time, even if sometimes the lights change for no reason, it is due to the inaccuracy of the RSSI.
We also tried one last thing : one chip is fixed, an other one is moveable and sends its RSSI to the location chip which can tell the distance between the two first. We taped a chip to the remote-controlled car and tried to find the hidden chip with the location chip next to us. It worked good, excepted for the delay before the location chip tells the distance with the lights and sound.
The location team,
Nicolas and Adrien