Cohesive Technology (yours truely) is creating systems and iterfaces specifically for motorsports.
I just finished a 2007 BMW K1200LT fitted with a dual-core Atom. It is on the road now with Ride Runner and DigitalFX (a temp UI).
Hi I have been developing a system to capture images of people and vehicles that tailgate my car.
The background of this is that I been actively working at reducing the amount of gasoline that I use per week in my daily commute to downtown Washington DC from suburban Maryland.
In 2004 I began monitoring my fuel use and found my round trip of 34 miles daily was requiring 11.3 gal/week. Next I joined many forums, www.ecomodders.com, became a member of the electric vehicle association of Washington DC http://www.evadc.org/. This is all to learn what I could about fuel economy.
My car is a 1996 VW Passat VR6 not exactly the most economical car out there.
However I have been able to reduce my fuel consumption to 8.1 gallons per week. The savings works out to 8.8 gallons each month. Meaning I don't have to pay for gas one week each month. With regular costing $2.80 it works out to almost $300.00 saved for the year.
The technique is called hyper milling, which means stay at the speed limit, use the terrain, use gravity down hills, coast up to stop lights, low roll resistance tires, high tire pressure and listen to mellow music that other stuff makes you drive fast.
Enough background, hyper milling comes at a cost. When you only do the speed limit other drivers want you out of the way. Hence they tailgate, which I don't really mind as the facial expressions are hilarious.
I created this website http://www.didyoutailgateme.com/
But I need a system to help me with this so I devised this system. It consists of a range finder, camera to capture license plates, camera to image drivers, gps for speed monitoring. And a lot of C++ code.
Things I have done to this point:
Face detection, here I use opencv libraries which are in C++, it can find faces in live feeds and lets me capture to jpeg's. My code packages those capabilities.
Range detection, I had hoped to use my backup sensors but the range is limited to about 7 feet. On a robotics forum I found a MAX Sonar which ranges out to 25 feet. I wrote a C++ dll to work with that.
GPS for speed, location and distance traveled while being tailgated. I have a Garmin GPS-18 and iguidance but I wrote a C++ GPS parser into a dll that my system can use.
I have a 8-port relay board that I wrote a dll for to control selection of devices when tailgaters are detected.
My next step is to make time to put at all the subsystems into one master program that runs in the background. I would like to write a skin for RR/DFX to allow customizations on the fly. And soon I will be able to start posting photos of tailgaters onto my website.
I would use the grant money to purchase an elm odbc device to the carputer. I want to make use of Google earth to interpret the terrain in front of the car and give acceleration/deceleration advice depending on the projected topography.
Here is a link to shoots of my car http://www.cardomain.com/ride/2742632
I believe that would pass any litmus test as to benefit to the community.
Although my idea of benefit to the community is a much larger community as you can see.
I have no love for hipermilers and I don't see any benefit on this.
I don't mind they playing with this in a deserted road, but they can be really annoying in traffic. Some of them have no respect for the others who just want to follow the flux. And yes, I've seen a lot here in the Bay Area, specially around the campus.
(Autoblog green)This is fine when a vehicle is driving in little or no traffic. Unfortunately, when there are other cars around, an extreme hypermiler can pose a real impediment. Differences in speed are often a source of accidents and people who aren't driving hybrids won't want to be stuck behind someone coasting along. There is nothing wrong with trying to maximize fuel efficiency and everyone should be trying to drive efficiently. However, even hypermilers need to be cognizant of other drivers around them and avoid being obstructive.
As you can see, people over react to the speed differential and that causes a traffic jam where there would be no need for one in the first place.
In traffic, they can be a real impediment to the flow and cause more waste, because everybody elses either looses momentum or time.
So I vote no on this one, but that's me.
Presumably, the ELM device, which is not particularly expensive, could help this project along and make for an interesting report on using car pcs to hypermile. Technical details of the project would be nice in a forum worklog and might attract hypermilers to our community, spurring additional innovation. Or maybe not.
Whether the project benefits specific members or not, the idea is to encourage innovation. I, for example, have no specific need for voice control of my PC but others do and a grant was awarded to a member for that purpose. The idea of an innovation grant is to expose the community to different areas of development that encourage overall innovation. If that leads to the development of a specific solution that they find useful, great. But it isn't a prerequisite.
My OLD 2001 Mitsubishi Eclipse GT:
"The Project That Never Ended, until it did"
next project? subaru brz