Does Prius Draft Exceptionally Well?

Having an automated system improves the, automatic reaction time. In theory and what makes sense would be three or four Prius that like a trained team of bicyclists would 'trade the lead' every 10-15 minutes. The advantage would be the lead, pacing vehicle task would be shared but switching leads would be tricky. It would take car-to-car communications and coordination.

Given our USA traffic patterns, the approach that makes sense would be if the tail-end car passes the others and pulls in to become the new 'front' location. Then the new second car enables DRCC to sync with the new lead. Roughly every 10-15 minutes, the lead would change. The team model would be:

• "A" MPG - mileage of single vehicle at given speed
• "B" MPG - mileage of following vehicle at same speed
• team MPG = (1/n)*A + (n-1/n)*B

It would help if you might someday run MPG tests both following and solo over a given course with another ZVW30. This would allow quantification of "A" and "B". Since we have data points for 55, 65 and 75 mph, these would be ideal data points to gather:


There is a 'wild card' that has to do with dampening in a train of Prius using DRCC. What might happen is DRCC speed trimming will have a certain amount of 'slop' that tends to increase in magnitude the further down the 'train' of cars. The trailing car would see the greatest variation in speed as they 'ripple' down the chain. What we don't know is the dynamic characteristics of DRCC and whether there are speeds and settings that are stable, dampening these variations, or unstable.

Forming DRCC 'trains' is not a technique for getting insanely great mileage but a way to cover great distances at better mileage. It would be a great way to cross Kansas. But first we need some metrics.

Bob Wilson

 

"Never to be run experiment"!!!

Sign me up! Let's get a few more Prius V w/ATP together and we'll do our own Lincoln Highway shuffle.

Anyone out there game? I'm sure Bob will be happy to supply the plan for us and analyze the data we gather.

 

Indeed I would!

Step 1 - identify Prius V w/ATP home locations

It would be perfect if we could find three volunteers within easy distance of each other ... say 240-300 miles. The two outliers drive to the center volunteer on or the day before 'test day.'

For grins, what are the volunteer distances from Georgetown KY where Toyota has a plant that makes Camry hybrids? We could coordinate a plant tour on the holiday morning and run the tests on Sunday.

Step 2 - identify 'rally day'

Obviously Hybridfest would be a natural but there are sooner holiday's coming up:

• Columbus Day - Oct 12, 2009, Monday
• Veterans Day - Nov 11, 2009, Wednesday
• Thanksgiving / Christmas - Ok, let's be real
• New Years - Jan 1, 2010, Friday
• Martin Luther King Day - Jan 18, 2010, Monday
• Washington's Birthday - Feb 15, 2010, Monday

Step 3 - instrumentation and route planning

Based upon the starting location, a planning breakfast and a short out and back run ... say 1 hour out, break and compare notes, and 1 hour back. Lunch and look at the initial set of data.

Afternoon run say 1 hour out, break, compare notes and 1 hour back. Data analysis and collect notes.

Instrumentation:

• GPS mouse and laptops to record location and velocity data with 1 second resolution -- differential analysis of the data gives very high resolution metrics. (I have three laptops and can get two more GPS mice so all three cars are instrumented.)
• (optional) - WiFi network for vehicle-to-vehicle communication and live data analysis (Ok, so I may have too much time on my hands.)
• Trip meter recording - a basic technique, all it takes is a passenger or audio recording device (aka., iPod/iPhone.) The passenger could record the data on the laptop in each car.

Risk: rain although the route can mitigate this risk ...

Bob Wilson

 

Only problem is that I think you have to be a lot closer than the CC will allow to make this work. Of course this has been done many times. I think by Carnegie Mellon maybe?
Anyway I have seen on TV, science channel? Where there were 4 or 5 cars within a couple of feet of each other in an experiment being automatically driven, automatically driven car "trains" with different destinations, but would group up for mileage, and individuals would drop out and be added.
In aircraft there is what is called ground effect which is only effective at less than 1.5 times the wing span or rotor span, it's not linear and to get much benefit you need to be .5 or less.
Completely apples and oranges of course, but I think to get real benefits you will need to be within a meter or so of the car ahead of you. That's not really hard to do at all if traffic allowed it, I'd be willing to give it a go.
Oh and if my bicycling experience means anything, the more bicycles there were, the better the draft. In the middle of a 30 or so paceline at 30 MPH, you felt like you were almost coasting, whereas by myself at 30 MPH, it was a sprint speed, I could do it, barely, and not for long. Difference was that big.

 

One of the reasons I'd like to have two or three of my laptops with GPS mice and WiFi between the cars is to see how accurate and responsive differential GPS would be in this application. The theory is three GPS receivers from the same vendor will calculate the same 'error offset' in parallel. Their ground track will wander in proportion to the GPS induced error but all three should have pretty good accuracy between each other.

With a nice ad hoc WiFi network between the vehicles, streaming their GPS ephemeris along with VoIP for intra-car communications ... it might work. Initially, this would be 'man in the loop' systems but later ... well if you know where the accelerator and brake signals come from ....

Bob Wilson

 

Bob do you mean differential GPS or WAAS enabled GPS? Differential I think requires a stationary transmitter and has been replaced pretty much by WAAS unless extreme accuracy is required, I think.
Close formation flying isn't as hard as people think, motorcycle racing within a foot or so of each other isn't hard, I don't think driving within a meter or so would either, steady state of course. It would be a fun experiment anyway

 

Differential GPS. Since our interest is in relative x,y,z and dx, dy, dz, the 1 second updates from identical mice recorded to and WiFi shared between laptops should provide very accurate emphmeris.

I'll order a set of mice this week (aka., same vendor and model) and hack the software out next week. Now if someone were to have a 3-dimensional capable vehicle with a clear line of sight to a ground based laptop with WiFi ... <GRINS>

Do you prefer meters or feet? ... It does not matter, I'll provide both.

Bob Wilson