Weekly report: January 25 - Februry 8, 2018

Transcription

1 Weekly report: January 25 - Februry 8, 2018 Yerbol Aussat February 8, Activities Built four sensing modules on Onion Omega 2 machines Installed the firmware and all required libraries, and set up TCP socket servers that measure light level and motion Set up TCP/IP socket communication between four server Onion machines, and one client Onion machine Collected more Light level vs Power data for Homedepot DC bulbs, white Philips Hue bulbs and color Philips Hue bulbs Marking the assignments for CS-240 and working on the assignment, paper reviews and project proposal for CS-889 (Computer Vision) 2 What I learned (insights) 2.1 Illuminance vs Power vs Control Value measurements Figures 1-3 show the results of two experiments for the same Homedepot DC bulb. Figures 4-6 show the results of two experiments for the same White Philips Hue bulb. Figures 7-9 show the results of two experiments for the same Color Philips Hue bulb. 2.2 Standalone sensing module on Onion Omega2 Last week I installed the latest version of firmware on four standalone sensing modules on Onion machines using a command line as described in [1]. Before installing firmware, I changed IP address on each of the Onions in their AP networks as described in [2]. All new Onions have default IP addresses , therefore I changed them to unique ones to avoid future communication conflicts such as offending keys. 1

2 Figure 1: Illuminance vs Power for a DC bulb Figure 2: Illuminance vs PWM Duty Cycle for a DC bulb Figure 3: Power vs PWM Duty Cycle for a DC bulb 2

3 Figure 4: Illuminance vs Power for a white Hue bulb Figure 5: Illuminance vs Brightness for a white Hue bulb Figure 6: Power vs Brightness for a white Hue bulb 3

4 Figure 7: Illuminance vs Power for a color Hue bulb Figure 8: Illuminance vs Brightness for a color Hue bulb Figure 9: Power vs Brightness for a color Hue bulb 4

5 Next, python-light with python-enum34, OmegaExpansion and pyonioni2c modules were installed on each Onion machine. These are the libraries that were used in the script for interacting with the sensors. It should be noted, that the guide in [3] suggests to install the most recent pyonioni2c, which is broken. Therefore, I used the steps described in [4] to install the previous version of pyonioni2c library, namely, pyonioni2c v Analogue vs Digital AMN PIR sensors In the four sensing modules digital AMN PIR sensors were used. It should be noted that digital and analogue PIR sensors have different outputs. While analogue AMN PIR sensors return a value that is typically between 0 and 150, digital AMN PIR sensors use GPIO s, and therefore only capable of returning binary 0 and 1. The sensor returns 0 when no motion is detected, and 1 - when the motion is detected. Figure 10 shown the block diagram output circuits for both analogue AMN and digital AMN versions of the sensor. It can be seen that distinguishes the two sensors is the comparator output circuit built in the output of the digital sensor. diagrams.png Figure 10: Block diagram output circuits of digital nad analogue AMN PIR sensors [5] As shown in Figure 11 [6], the digital sensors take the amplified analog signal and convert it to a digital signal by using a preset threshold voltage programmed inside the ASIC. As the analog voltage goes above the positive or negative threshold values, a digital signal passes through the output of the sensor. This threshold cannot be directly adjusted by the sensors user. 5

6 Figure 11: Output signals of digital nad analogue AMN PIR sensors [6] Therefore, it can be seen that the analogue PIR sensor provides a much richer output information, compared to the binary output of its digital counterpart. This allows to use adaptive thresholding with the analogue AMN for occupancy detection, such as analyzing standard deviation of the last several readings. After some initial tests with the digital AMN sensor, many falsepositives have been seen. In other words, the digital sensor outputs 1, when there is no motion in the room. As it was suggested in [5], using pull-down resistor might help to alleviate this issue. I will try adding a pull-down resistor next. 2.4 Communication between sensing modules and the controller After trying to change the wifi network for the Onion machines, it was found out that using a console is not effective for this. Instead, the wifisetup command should be used in command line as discussed in [7]. I was able to successfully connect all of the sensing modules and the Raspberry Pi to the network of Omega-F109 machine (the one that controls PWM of the Homedepot bulbs). However, I was not able to make its children (RPi and 4 Onion sensing modults) communicate with each other. Connection timeout occurred every time I try to ping or ssh the devices from each other. However I found it strange because: I was still able to ssh Onion Omega from RPi and vice versa when both of them are in my local network at home. I am able to ssh to RPi and Onion sensing modules from my laptop, when my laptop, RPi, nad all Onion machines are in the Omega-F109 network. I am able to ping to RPi, but not Onion sensing modules from my laptop, when my laptop, RPi, nad all Onion machines are in the Omega-F109 network. 6

7 2.5 Communication between Onion sensing modules and the parent Onion machine I was able to successfully set up TCP/IP communication between the parent Onion machine, and four sensing modules. Four socket clients are set up on the parent Onion machine that request sensing data from four servers running on the Onion sensing modules. This works very robustly, and the parent Onion reliably receives an array of light level readings and an array of motion readings from the four sensing modules. 2.6 Controlling Philips Hue bulbs from Onion Omega using python script I was able to connect the parent Onion Omega to the Philips Hue s NETGEAR network by the means of the Ethernet expansion dock and the Ethernet cable. After installing Philips Hue python library on the parent Onion Omega as well as a number of dependencies, I was able to successfully control Philips Hue bulbs from the parent Onion Omega. 3 Next Try to improve the performance of digital AMN PIR sensors Make a binary control of Philips Hue bulbs from the parent Onion machine by taking into account sensor readings from four sensing modules. Try to setup wireless communication between Philips Hue bridge and other machines 4 References. [1] [2] [3] [4] [5] t o u se/index.jsp] [6] [7] 7