Fancy becoming a Beta Tester for a new motoring Android app?

A place for new members to introduce themselves and give some background information as to their experience with driving, likes, dislikes etc. It is advisable to post here first before posting to other forum's on the site.

Postby Gareth » Sat Oct 18, 2014 6:38 pm


DouglasPonsford wrote:In terms of advanced driving, I can say from experience (I passed the IAM test some years ago) that advanced driving is more than just being observant, observing speed limits, making progress, showing courtesy to other road users and following the rules of the road. It's also about keeping our vehicles in good shape and generally trying to minimise our impact in the wider environment. Having a method to be able to measure those aspects of good driving and generally having a mechanism to be able to see if there's room for improvement is also a way to advance anyone's driving skills.

I'm wondering what aspects of advanced driving can be measured. In this answer you seem to be focusing on economy, but I'd be more interested if there was some way to measure the combination of smoothness and forces acting on the vehicle, and then to be able to identify those aspects that relate to patterns of advanced driving.

So, for example, people learning advanced driving through the IAM or RoADAR often have smoothness as a goal, but mistakenly think that smooth braking means long drawn out braking, rather than smooth transitions onto and off the brakes interspersed with firm braking. The latter is better as StressedDave explains more lucidly, because light braking over a long distance doesn't help with planning what to do at the hazard. It's also interesting how braking is tapered, because doing most of the braking away from the hazard leaves more time for assessment as you close up on the hazard.

Together these make the following pattern that might be identifiable: the smooth introduction of braking over a shortish period leading up to a longer period of sustained firmer braking, then smoothly tapering off with an extended tail before stopping or accelerating.

Another example would be smoothness in steering - a smooth transition from straight ahead increasing until enough steering deflection has been achieved, maintaining that deflection for as long as is necessary, then another smooth transition back to straight ahead. Good steering would be proactive instead of reactive; the latter might be detected by a slight unwinding of steering lock once the vehicle is turning as much as the driver originally desired.

Yet another example might be to detect and record smooth yet firm acceleration up to a speed limit - missing out needless gear changes and block changing once the upper speed has been achieved. Given that the speed limit is often the limiting factor on how fast we drive, it ought to be possible to record how well we maintain the speed limit once it has been reached, versus whether we allow the speed to drift up and down, and by how much, IYSWIM.
there is only the road, nothing but the road ...
Gareth
 
Posts: 3604
Joined: Sat Apr 15, 2006 2:58 pm
Location: Berkshire




Postby DouglasPonsford » Tue Oct 21, 2014 6:32 pm


Gareth wrote:I'm wondering what aspects of advanced driving can be measured. In this answer you seem to be focusing on economy, but I'd be more interested if there was some way to measure the combination of smoothness and forces acting on the vehicle, and then to be able to identify those aspects that relate to patterns of advanced driving.


Yes, I agree totally. This is an aspect of the application that we are developing, alongside the economy aspects. We expect the majority of our customers to be 'ordinary' drivers, who will be interested in not just getting from A to B, but also how to squeeze more economy from their driving and also may be interested in learning about techniques that will improve the driving experience for themselves and their passengers (and also extend the life of some of the components of their vehicles). Most of our customers would not be considered 'Advanced' drivers however and economy is going to be one of the major factors influencing most non-corporate users' decision to sign up for our service.

Having said that, it's fair to say, of course, that the techniques used in advanced driving overlap to a great extent with the techniques required in driving for economy; advanced driving usually goes hand-in-hand with economical driving, so whilst we are not trying to develop an application that will specifically teach people advanced driving techniques - and neither do we claim to do so - some of the tips and advice that we will provide to our users for better economy will also be the same advanced driving tips that you mention.

The smooth driving techniques that you describe for braking, acceleration and steering for example, all follow a fairly well defined profile. If vehicle deceleration, acceleration, steering input, braking pressure and throttle position are measurable, then the applied profiles of these metrics can be compared with idealised driving profiles. Allowing a tolerance for outside influences like pedestrians and cyclists with a death-wish, we can start to apply figures as to how near or how far an individual driver might be to this ideal profile.

The hard part, as you may already have guessed, is getting access to that data in a form that allows measurement and comparison of a level of the driving 'ideal'.

The primary hook into vehicle data is through the OBD port. Of course, this does generally limit the full benefit to vehicles built since 2000, but does still cover the majority of vehicles in use in the UK today. Vehicle speed and throttle position are generally available this way but only a few high-end luxury vehicles provide data about steering and braking input (very few of the Japanese types, except for hybrids, provide much in the way of data at all, although this is getting better with newer models). None of them provide g.

With vehicle speed and throttle position, we can get a good estimate of accels and decels, but only in a straight line. Yaw moments and large steering inputs cannot be detected this way, so we need accelerometer data. The problem here, is that we don't (yet) have a reliable source of accelerometer data. We could try to use the g sensors in the phone, but this relies on the phone being fixed in relation to the vehicle, which we can't guarantee (although we have a method for dealing with that in the pipeline).

As for constant speed driving, this would really only apply out of town, and particularly on motorway scenarios, so although we could detect motorway driving and add in a constant speed/constant throttle factor, this would be of less use in town, as there are so many variables outside of our control (crossings, traffic lights, congestion, roundabouts, idiots on cycles etc), it would be difficult to maintain a constant speed in order to be able to reliably calculate any deviation from 'ideal' driving behaviour.
DouglasPonsford
 
Posts: 8
Joined: Thu Oct 09, 2014 7:26 pm

Previous

Return to Introductions

Who is online

Users browsing this forum: No registered users and 1 guest


cron