This week, we worked on getting a demo working and started working on integration. The demo was integrating the ESP, pinging, and localization to get a sense of how accurate the localization algorithm is. We decided to not use PicoScenes as we were able to get data we need from the ESPs, so some of what I’ve been doing is no longer useful, but it is good that we can get all necessary data from the ESPs since I wasn’t able to figure out a way to make PicoScenes detect the ACK packets from our pinging. Meanwhile, I worked on testing and cleaning up the device detection script, which it does work. I’ve changed from manually sniffing beacon frames to using iw for the access points, since manually sniffing can sniff access points from adjacent channels which show up as two same access points on different channels. I’ve also tested how long it takes for the devices to respond to my beacon packets, and it seems like about 7 seconds is enough. This can’t get too long, as there may be arbitrary number of access points and our aim is to have everything (from device detection to localization to visualization) be within 5 minutes.
I am testing my part on my local network, and have verified that it detects my phone and ipad consistently. More tests would be possible when our device arrives, with multiple access points that we have control of. Next week, we will continuously work on integration, and I will personally look at device identification to show the users, such as manufacturer, the transmission rate, etc. I am a little bit behind on the device identification, but it should be possible to catch up by next week.