JimBob
Well-Known Member
- First Name
- James
- Joined
- Jan 26, 2021
- Threads
- 75
- Messages
- 950
- Reaction score
- 1,109
- Location
- Toronto Canada
- Vehicles
- Taycan 4S
- Thread starter
- #1
We all use the Range Estimates as calculated by the car, but are they accurate?
Using the data from my recent cold weather test, I looked at the error between the estimated range and the actual range. The Total distance covered was 338km/210.1m Observations were recorded at 2 minute intervals yielding 83 discrete data points for the trip, equivalent to 83 forecasts by the Guessometer.
The results are shown in two quick and dirty scatter plots. The first for metric users and the second for Imperial users. Each observation is 2 minutes apart from its neighbor. A rough interpretation is that the Guessometer starts out inaccurate and then converges to the point where there is little error.
Observations:
The car attempts to forecast the range based on past behavior
Initially this is not very accurate. The plots show a large relative error initially which eventually converges to the actual range
Each person will have their own estimate as to what is an acceptable miss and how long the miss persists until corrected.
The car was put on cruise control as soon as practicable
This is probably the easiest calculation for the Guessometer to perform. Constant speed with few outside interferences
Cruise control was initiated within the first 10 minutes of departure and then maintained as long as conditions permitted
In this case mostly until the half way point out and then most of the way back until traffic required it to be disengaged
If the Guessometer were perfectly accurate the Range Error would quickly decline to 0
The Guessometer, at constant velocity became extremely accurate after 60 minutes
Probably pretty accurate after 40 minutes
Once the Guessometer settled in and speed remained constant, it remained very accurate
The decline to zero at the end is because the test was terminated at 10% battery charge and the estimate became the range
Comments
The initial error could easily be corrected by Porsche with what should be an easy software fix. The car knows most of the parameters but it doesn't know the speed at which the driver intends to travel and it might not know the outside temperature. If it let the driver input an expected speed and air temperature
it could start out quite a bit more accurate. If the drivers estimate turned out wrong, the Guessometer would eventually correct the error.
If it was implemented in the APP on your mobile phone, the driver could test for travel at different speeds from the comfort of his favorite chair. I think this would be useful.
Using the data from my recent cold weather test, I looked at the error between the estimated range and the actual range. The Total distance covered was 338km/210.1m Observations were recorded at 2 minute intervals yielding 83 discrete data points for the trip, equivalent to 83 forecasts by the Guessometer.
The results are shown in two quick and dirty scatter plots. The first for metric users and the second for Imperial users. Each observation is 2 minutes apart from its neighbor. A rough interpretation is that the Guessometer starts out inaccurate and then converges to the point where there is little error.
Observations:
The car attempts to forecast the range based on past behavior
Initially this is not very accurate. The plots show a large relative error initially which eventually converges to the actual range
Each person will have their own estimate as to what is an acceptable miss and how long the miss persists until corrected.
The car was put on cruise control as soon as practicable
This is probably the easiest calculation for the Guessometer to perform. Constant speed with few outside interferences
Cruise control was initiated within the first 10 minutes of departure and then maintained as long as conditions permitted
In this case mostly until the half way point out and then most of the way back until traffic required it to be disengaged
If the Guessometer were perfectly accurate the Range Error would quickly decline to 0
The Guessometer, at constant velocity became extremely accurate after 60 minutes
Probably pretty accurate after 40 minutes
Once the Guessometer settled in and speed remained constant, it remained very accurate
The decline to zero at the end is because the test was terminated at 10% battery charge and the estimate became the range
Comments
The initial error could easily be corrected by Porsche with what should be an easy software fix. The car knows most of the parameters but it doesn't know the speed at which the driver intends to travel and it might not know the outside temperature. If it let the driver input an expected speed and air temperature
it could start out quite a bit more accurate. If the drivers estimate turned out wrong, the Guessometer would eventually correct the error.
If it was implemented in the APP on your mobile phone, the driver could test for travel at different speeds from the comfort of his favorite chair. I think this would be useful.
Sponsored
