Introduction
|
|
Project: Watchtower Introduction |
|
|
|
|
|
|
|
|
|
Everybody has a boss button defined on the keyboard in the office. The problem is to press the button in time, especially if you have as sneaky a boss as mine. I have put up a mirror on my cubicle wall to catch her creeping up on me, but in order to justify it, I have to go to great lengths to show that I take care of myself by putting hand lotion, chap-stick and a hair brush on the table too. To get around this, our ‘eye in the sky’ is being proposed. Well, sort of eye near the ceiling of the office. We call it the Watchtower. It hovers above the cubicle walls obstructing your view. The video image of the boss is relayed to you, which you can see on your video screen, and press the boss key with plenty of time to spare. This project is a lifeline to those who believe in the high principles of playing video games in their cubicles, and have bosses who just won't listen to the voice of reason. It's not just a sneak-o-scope ----- It's a SNEAK-O-SKOPE WITH A PURPOSE ------- Project Milestones:
NOTE: Milestones already achieved are in blue. Progress Notes: 12/31/2009: All the hardware has been procured, including material for air frame. The SBC has been chosen, though that may change if the project goes through round 1 of Spark 2010 Challange. The airframe has been designed, and in the process of being finished. The sensors and actuators have been tested. Currently, work is in progress on OS Design on Windows Embedded CE 6.0. 1/1/2010: Tried installing Adruino IDE on a VMWare image. Did not work. Will try to install it on a separate machine. The development environment is based on GNU toolchain. 1/2/2010: Given up on VT6655 Mini PCI WiFi card. Works fine on Windows XP, so nothing wrong with the hardware. Will proceed with Linksys bridge, with which earlier bench test was done. 1/4/2010: Bought the VoIP wireless adapter (see Implementation page for details). The saving in weight is significant. Without the casing and the wire, the board itself is just 15 grams. The wire is 16 grams, and I think it can be shortened further to reduce the weight. 1/9/2010: Configured Arduino IDE and development environment. Also created a headless image OS Design in Windows CE 6.0 for ebox 2300SXC-A box. If the project goes through the first phase, then DMP is already providing headless image solution and sample application for Vortex86DX processor, and the project will be moved to the competition hardware platform without any problem. 1/17/2010: The project didn't go through Round 1 of the Contest. Therefore, the original goals have changed somewhat. Since only 300MHz of processing power will be available instead of 800MHz on the contest board, CPU cycles have to be budgeted very carefully. On the positive side, the project does not have to adhere to the contest rules of putting most of the processing on on the SBC. Atmega328 microcontroller had a lot of processing cycles left over, so it was decided that Kalman filters will be implemented at the front end on Arduino board itself. Very simple single stage Kalman filters have been implemented. If in flight testing, the controller performance is not good, then the option of upgrading SBC to a more powerful Vortex86DX based SBC will be made, where more a complicated extended Kalman filter can be implemented. 1/25/2010: Platform builder on Windows 7 does not run very smooth. Tried installing monthly updates, but I think I haven't installed the updates for both VS2005 and PB in the required sequence, because of installation of updates failed. In the end, there were two options left. Either build a machine with Windows XP (VMWare, as earlier pointed out, doesn't work), or move to Linux. The first option meant more budget, so that option was ruled out. In the end, installed Linux kernel 2.6.24 on Ebox 2300SX. Everything tested to be running fine. Next week, I'll configure wireless link between base station and vehicle controller (2300SX). Both the Vonets bridge and the Linksys bridge gave me a lot of grief while connecting with Verizon Wireless router with 64-bit encryption. I'll connect it with dedicated Linksys Router with no encryption to see if that works or not. 1/26/2010: Connected Vonets wireless bridge to Linksys router without encryption. Worked great. Next weekend, I'll put all the components together, and write a simple application to test the interfaces. 1/31/2010: In a combination of rotor-tower/iron-bird/strapped-down-bird test, the fuselage fell apart. It was re-fabricated with more enforcements, which resulted in higher weight. But hopefully, it won't just break into pieces mid-flight. 2/7/2010: Put Avionics components together, and tested all the components on the vehicle itself. The platform is ready now, and software development will start from next week. 2/19/2010: Unfortunately, I have to put the project on hold until summer. The basic software framework is in place, but the controller is still not implemented on the SBC. Processed signals from the sensors (IMU and GPS) are being routed to the Ground Station, and control signals are being sent directly from the Ground Station to the actuators. SBC software has been designed in such a way that controller can be implemented on it later, and only navigation and higher level control will remain on the Ground Station. All the interfaces that needed to be designed on SBC are operational. That's it, then, until summer. The software design, as it exists today, is given in implementation section.
|
|
|
|
|
|
|
