We can split up the architecture into 3 major blocks of abstraction. The lowest level would be the sensors connected to a BLE chip and small power source. On the mid-tier, we would have aggregator modules that would aggregate the chirping from the BLE-enabled sensors and push them onto a server via Wifi. Lastly the server would amass the information from the various aggregators, process the information over time and display it to a user interface such as a website or a phone application.
For the bottom tier, we would be using a variety of sensors to collect data from multiple sources and for different purposes. We have the following areas in mind:
1- Monitor ingress and egress points with KINECT sensors to get an accurate account of how many people are in a specific room or facility. This can be used to verify and check against how full the subsections are in that areas and as an indicator of overall capacity load.
2- Monitor individual assets such as chairs in the quiet study areas or in an area where chairs can be shifted around using BLE-enabled accelerometers. This allows us to monitor areas which can change dynamically over time without having to be incredibly specific to the details of the setup. The main goal with these sensors is to properly duty cycle them such that we can ration the power efficiently, having the sensors either wake up on activation or wake up periodically.
For the middle tier, we place aggregator modules in sections around the BLE-enabled sensors, so as to gather the information from the area into a single node. This module will also have an Wifi chip, through which it will relay the information to the server. The main rationale for this design decision was to allow the sensors to be aggressively duty cycled to reduce power consumption, and instead have these aggregator modules listen in on the bluetooth signals. Hence the power consumption for these modules will be higher, but that will be offset by the lower number of these modules. Additionally, these can also be powered from an outlet if necessary. This will reduce the long term maintenance of replacing the power supply for the numerous sensors.
Lastly, for the final tier, we aim to have a server running the software side of this project, which will sift through the data loggings and update the frontend website as required. We aim to have this update periodically, at a time interval of about 15 minutes.
Name: IC ACCELEROMETER 3AXIS 16QFN
Name: WiFi Module - ESP8266
Memory: 1MB Flash
SimpleLink Bluetooth Smart and Proprietary Wireless MCU
Vendor: Texas Instruments
Name: Raspberry Pi 2 Model B
Name: BATTERY LITHIUM 3V COIN 12.5MM
Size: 12.5 mm
Coin Cell Holder
Name: HOLDER BATT PC COIN 1CELL