Results

Wireless Ambient Lighting Control System

Pratik Panchal (pp423@cornell.edu), Kedar Vidvans (knv25@cornell.edu)
M.Eng, School of Electrical and Computer Engineering,
Cornell University 

High Level Design  |  Hardware Design  |  Software Design  |  Results  |  General Issues  |  Appendix  |  My HomePage


Results and Conclusion:

We got precision light control. The PIR sensor also worked well to conserve energy in case it detected no presence in the vicinity. Our accuracy for control was within a couple of PWM cycles as defined by the modifiable variable PWM steps.

Initially we had bought wireless modules from sparkfun. We faced difficulty in implementing the implementation of wireless protocol. We started by using ASK transmitter and receiver modules by spark fun working at 315 Mhz. But they did not work no matter what we tried. We tried different methods tried by previous years groups like sending syncronizing byte (0xAA) multiple number of times before the actual data was sent to tune the receiver, sending balanced bytes (equal number of 1’s and 0’s), sending checksums, but all the above methods did not give a satisfactory result. We believe this might have been due to noise or faulty receiver-transmitter (caution to all future users of this product: Don’t use it, and if you do, dont be afraid to change it).

The tuning of the control loop was done with a  wired connection between the sensor unit and the control unit. The control strategy we had adopted was a feedback control, hence the sensor feedback could not have any delay otherwise the controller remain  “hunting” (oscillatory output). Since this the sensor data was not continuous and we had to account for dropped packets, the control loop was tuned to run every 30 msecs. This way we could also overcome the problem of spurious measurements and intermittent blocking of the sensor. 

The LEDs and sensor are not perfectly linear elements, but since, we designed the system for perfectly linear elements, at low and high light intensities, the LEDs would “appear“ not to give the same bang per buck. For a commercial product, one would have to implement a non-linear (degree 2 / 3) characteristics to get an accurate reading from the sensor.

The PIR sensor worked well, but, it had to sense continuous motion. This would be of disadvantage and switch off the lights, in case a person remained stationary for a long period of time. Also the sensor has a limited range of 4-5 meters. Hence for a large room areas, one would want sensors with higher range.

We were demonstrating this for a single sensor unit and light source, realistically, one would want more that one sensors, all placed near sources of light and work area. the control unit would then be able to control the light sources independently depending on natural sources of light and the location of humans in the room. 

Due to budget limitation, the light sensor and PIR are positioned on the sensor unit and control unit respectively. Due to this, their “view” is blocked by wires and other components. Ideally you would want these to be in location of minimum obstruction.

The range of the wireless is more than 25. (we checked diagonally across the lab.)


Disclaimer: This work has been done as term work for the course ECE4760 Digital System Design using Microcontroller, at school of Electrical and Computer Engineering, Cornell University by Pratik Panchal and Kedar Vidvans under guidance of Prof. Bruce Land. Readers can use the presented work on this site as long as they acknowledge the source. The work is presented in as-is condition and If used, no liability is borne by either the authors or the school.


web counter
Comments