Advanced Eye-Strategy in Highly-Skilled Race Driving – a Review
Quite often the advice to drivers is to ‘look further ahead’ or ‘look at the exit before you hit the apex’. The base concept is ‘look where you want to go’ and this will result in forcing the brain into ‘flow state’. While this might be useful and produce results, it doesn’t really get to the heart of what is really happening on a biological and neurological level. The whole education of looking ahead is based upon replication of observed behaviors in good drivers and not much more. Though we are starting to see more publicly accessible studies look into the phenomena of race driver eye strategies, we are still some way from a comprehensive analysis.
The objective with this article is to nail down the basics of eye movement in driving so we can start to build a solid foundation for eye strategy. One thing to note is – rarely do people know where they actually look. Eye trackers, in all sorts of different kinds of research, quite often demonstrate that people don’t look where they think they are looking. It’s quite remarkable and reminds us that when someone tells us where they look in a race car, without hard evidence, it has to be treated as theory, not fact.
So let’s get started and figure all this out. I am going to go through some old studies and some new ones with a dose of anecdotal experience so we can deconstruct and try to understand advanced eye strategy for race drivers.
Basics of Eye Movement
So to begin, there are two main, though not exclusively, types of eye movement – saccadic and smooth pursuit. We will concern ourselves mainly with saccadic movement as that’s the one we see most in racing environments.
Saccadic movement is, in basic terms, a series of short fixations separated by fast movements of the eye called ‘saccades’ that last around 30ms. If you have ever observed someone’s eyes closely you will soon notice the eye constantly having small fixations then moving. This is seen with eye trackers on race drivers too.
Saccades are generally pre-programmed and thus any change in target after the movement has begun can not change the destination of the fixation. During the saccadic movement no information is being recorded. As saccades make up around 10-15% of our daily time, that means for about 1-2 hours a day we are blind, or to put it another way, blinks not included, in a race that lasts about 1 hour you spend 6-10 minutes totally blind. Crazy, right?
Another quirk of saccadic eye movements is they are not 100% accurate, and this may have implications if you have a strategy for reference points while driving. Typically, for any saccadic movement that is over 20 degrees the eye will make a small correction of around 10%. These higher amplitude movements are followed by a correction saccade that can have an interval of 150ms which is a tenth and a half. That’s a lot in racing terms.
So there we have a very basic understanding of saccadic eye movement, something we do thousands of times during a race.
Anticipating the car – the Reflex systems
You feel it through seat of your pants, right? Well, how we feel a car’s dynamic as it moves through space is somewhat a mystery from a scientific perspective. It’s like magic. When entering and going through a corner we are anticipating movement, while reacting to various other nuances of car behaviour in areas of the brain not yet explored. It’s truly remarkable how our brains are adaptable enough to accept this new reality. A reality which has very little comparison in the natural world.
We have a number of reflex systems but there are a couple worth looking at and understanding as a couple of interesting things were observed in research.
Vestibular Collic Reflex (VCR) – The reflex that opposes the direction of the head in space sense by the vestibular networks of the inner ear. This means your trunk may change angle, but your head and gaze counter rotates your neck muscles so you can remain focused on a particular object.
Cervicollic Reflex (CCR) – A reflex that opposes the direction of the head in space, same as VCR, but is a result of neck proprioceptive inputs during motion of the body.
These reflexes are relevant because Grasso et al (1) discovered something remarkable about how we anticipate future direction change when we walk:
“Head and eyes systematically deviated toward the future direction of the curved trajectory. Anticipation lead was about 1s. Strikingly, the same behaviour was observed in darkness. In backward (BW) locomotion along the trajectory (from end- to start-point), gaze deviated toward the opposite direction, such that the forward locomotor pattern did not appear time-reversed. Orienting movements displayed higher amplitude, reproducibility and time lead in the forward (FW) direction at light. We suggest that anticipatory orienting synergies belong to the behavioural repertoire of human navigation and may reflect the need to prepare a stable reference frame for intended action”
So basically our heads anticipate the future change in direction very accurately around 1s prior to the body (also known as trunk) moving in the same direction when walking. It will read as something quite obvious at first, but it’s a remarkable discovery. It gets even more remarkable once you consider the same phenomena was observed in the racing driver Tomas Scheckter.
Land and Tatler, 2001 (2), found that ‘the head in yaw plane’ was an almost exact predictor of the rotational speed of the car 1 second later’ when analysing data from the experiment they performed at Mallory Park with Schekter.
Furthermore Peter M. van Leeuwen et al in their study (4) ‘Differences between racing and non-racing drivers: A simulator study using eye-tracking’ found that race drivers make greater head movements during turns than non-racing drivers. The researchers did highlight that the simulator provides greater movement of the head that may be found in reality with helmets and HANS devices, however it does confirm the notable significance of head movement in race driving.
We know in walking tasks the head in yaw plane is a predictor of future direction change about 1s prior to said direction change. So we can theorise that we use this same anticipatory head movement in a race driving scenario. The reason this strategy may not be used by non-racing drivers is lack of familiarity with the driving task, and thus have not developed the necessary anticipatory systems.
The race driver’s head strategy does come with a unique eye strategy which we will discuss in detail later, but if we take the head movement in isolation it could explain how some drivers interpret car behavior. Let me explain, if we move our head 50° 1 second prior to a turn yet the car only rotates 40° then we are experiencing understeer. If the car rotates faster than what we anticipated then we are feeling oversteer, this is possibly as a function of our VCR and CCR systems. This may or may not be the case, at this stage it’s just a theory, but it does allow us to observe high-level race drivers in a new light and analyse their body vs head strategies in real cars.
Certainly a driver like Alonso has a particular strategy in a car when he drives. The way he positions his body and head through corners is particularly unique. If it is the case that these collic reflexes provide us with information, then it makes sense the importance of core/trunk stability to avoid ‘bad feedback’. Lewis Hamilton even stressed how he used his ‘core’ to feel what the car was doing:
“The car steps out, you are using your whole core as well to feel it. It’s not like sitting there and reacting to the steering” Lewis Hamilton.
Max Verstappen is also another driver who has a very solid and stable core body position when he drives, most observable when he is in a road car. This is an area that clearly has huge potential for further understanding and something we will further develop over time.
It is worthy of note however that a racer can drive simulators at extremely high levels with effectively just the ocular systems and small 17in square screen. The wonders of the human brain mean that we may be effectively simulating the rotational forces in our brains to make up for the lack of feel elsewhere. It may also help explain why some drivers get sickness on the simulator as their eyes and reflex systems anticipatory nature conflict with the lack of fluid motion in the inner ear.
Science of ‘Looking Ahead’
Humans are able to navigate a race car at speed through a variety of curvatures and inclines. We seem to take this very complex activity for granted. Where we look provides us with the information we use to navigate and steer a car through a particular corner or complex, so it’s vitally important to try and understand as fully as possible. Though in the future we will discuss the difference between ‘seeing’ and ‘looking’ and investigate whether eye strategy is an actual effect of other functions in the brain.
It is widely accepted that very good drivers tend to look further down the road than average race drivers, though this is not conclusive. This isn’t unique to race drivers as it has been observed with road users (M.F. Land and C.J. Hughes, in Land, 2006) as well. So why exactly do we look ahead and what information is it giving us?
Humans tend to have a .8-1 second buffer where we fixate our gaze ahead of time before reaching said point/goal. In driving, this can be up to 2 seconds as observed by Edmund Donges 1978 (3). So for example our head vs trunk angle will rotate to its desired trajectory in a walking scenario before taking a bend about 1 second before our body rotates through the bend to come back in line with the head (VCR). So there’s very much an anticipatory nature to our gaze and where we look.
In very basic normal road driving terms what looking ahead does provide us with is information about the upcoming bend – its curvature and gradient. We can then make adjustments accordingly to navigate through the bend. In essence we are talking about a feed-forward system. However, looking ahead isn’t absolutely everything.
Michael Land’s discovery in 1998 demonstrated the importance of different areas of the scene we are presented with when driving. Working on a simulator (it should be noted simulated analysis of eye movement is not always correlated exactly with reality) he set about working out the reasons we focus in certain areas with our gaze during driving.
Land’s experiment was to analyse the importance of different regions of focus during driving. So the below illustration shows three regions. We have region A, which is at the tangent point and furthest away, we have B which is in the middle region, and C which is directly in front of us (fig.1).
Fig.1 – Land’s multiple regions for fixations during curve driving. Illustration – GTS RS
What Land discovered is that if we blocked region B and C the driver’s curvature matching was still very good but lane control was poor. Conversely if region A and B were blocked lane control was very good, but curvature matching was poor and lead to staccato steering inputs (not good for race driving).
Interestingly though, if only region B was blocked and only A and C visible, the performance was almost indistinguishable from when B wasn’t blocked and you had the whole scene. Having only region B visible provided a good result but was below that of combining A and C. If we consider focusing on the region A as a feed-forward system, the peripheral nature of region C may be considered a feedback mechanism. Combining these seemed to produce maximum performance.
Tangents and Apexes! Why do we look where we look
Where do we actually look and focus on when we drive and why? There has been quite extensive research done with road drivers which does allow us to garner a good understanding of steering input and the eye strategy of a driver on a normal road. Thanks to Peter M. van Leeuwen et al we can start to unpack what racing drivers do differently. Unsurprisingly there does seem to be a clear correlation between where a road and non-racing driver looks and their steering input.
Several studies suggest that road and drivers employ a strategy of tangential fixation and this was found to be the case with non-racing drivers too.
Tangent – a straight line or plane that touches a curve or curved surface at a point, but if extended does not cross it at that point.
The tangent point is the non-stationary point on a bend where the driver’s line of sight is tangential to the inner edge of the road – Michael Land & Benjamin Tatler, Steering with the Head.
What is advantageous about focusing on a tangent point of a bend is that it gives us the curvature of the bend. (fig.2) With some maths, which takes into account the car’s direction, the gaze of the driver (which is usually at the tangent point of the bend) we get an accurate representation of the curvature of the bend. This potentially allows us to steer accordingly along that path. It sounds obvious, but this theory wasn’t well established until the mid-90s.
Fig.2 – Gaze Direction and Tangent Point allows us to accurately navigate a curve. Illustration – GTS RS
Race Drivers and the Returnative Tangent Point Fixation Strategy
Why it is not wholly uncommon in beginner-level race drivers, the tangent fixation strategy is best used to predict the curvature of bends not maximise speed through them. You may even notice yourself doing this when you are learning the track for the first time (though humans are notoriously poor at anecdotally describing where they look, as eye trackers often demonstrate).
This strategy may have benefits in the building phase of learning a new track and it has been noted that regions of the brain dealing with novo environmental learning are strongly correlated with relative performance between drivers. So it should not be dismissed. Eye strategy doesn’t have to be one set rule. We can use eye strategy in varying ways to improve the speed that we learn race tracks. Looking 500m down the road before you fully know the environment may not always be the best way.
What Peter M. van Leeuwen et al discovered that race drivers tend to employ a strategy of multiple fixations. They state “As the racing drivers entered the corner, they
directed their gaze away from the tangent point towards the outside of the corner, and as they
progressed through the corner they moved their gaze towards the tangent point and beyond
the tangent point. As the racing drivers exited the corner, they directed their gaze again
towards the outside of the corner and subsequently looked again towards the tangent point”. This was particularly noticeable during Turn 1 of Mallory Park which the researchers focussed on as they replicated the study by Michael Land in 2001.
The limitation of the study was that the drivers weren’t given sufficient time to get to the absolute limit whereby the employment of different eye strategies may have been found. But what we are observing in race drivers seems to be a complex array of orientation and anticipatory movements of the eye. As mentioned already, it needs to be highlighted that when you try to apply looking ahead techniques in race driving, where you think you are looking may not be the reality of what’s happening. The brain is rather tricky like that.
The study speculated further that “These results are complementary to the results of a study on a real racetrack, which showed that a racing driver directed his gaze to the vicinity of the tangent point instead of at the tangent point. Contrary to models of visual control of steering in which reference points (e.g. the tangent point) are used to guide the steering input, our study shows that racing drivers vary their gaze while cornering, possibly to verify their path and to anticipate future control actions. Furthermore, the non-racing drivers may simply be looking at where they want to go, whereas the racing drivers may be directing their eyes to task-relevant information.”
What’s worthy of note here is when fixating back to where the tangent point is, race drivers fixation points were not entirely accurate, they were in the ‘vicinity’ of the tangent point. The researchers suggest that it may be a case of fixating where the drivers wants to go. However, if we go back earlier in this article we know that eye movements beyond 20 degrees tend to be inaccurate and require secondary correction saccade. It may well be the case that under stressful high-speed driving,accuracy of movement back to the tangent point of which is a continually moving flow field is of low priority even at lower degrees (<20) of movement. A secondary saccade takes time and could mean that the anticipatory function of looking further ahead is diminished in effectiveness.
Eye-strategy methods are often taught whereby race drivers are encouraged to eliminate Returnative Tangent Point Fixation (RTPF) in preference for a continuative ‘further ahead’ fixation strategy. A continuative fixation eye strategy may suggest a driver look at the apex before they hit it, then look towards the exit as they approach and pass the apex. A key feature would be to emphasise not returning gaze to the tangent region of the turn once the exit point is fixated upon. It may well be the case that in a situation where the driver is more familiar with the environment RTPF goes away naturally, and a more traditional look-ahead strategy will be employed.
According to Peter M. van Leeuwen et al, RTPF would appear to be the accepted strategy for highly-skilled racing drivers in the conditions of limited track time and an unfamiliar environment in pursuit of maximum performance. RTPF was especially prevalent during long corners. Therefore RTPF should not be discounted as a valid strategy during the learning phase of a particular track.
Michael Land and Ben Tatler’s excellent ‘Steering with the Head’ research suggested that with big sweeping turns for example tangential fixation is more likely as it remains there for a number of seconds, but on tighter corners the tangent point is more ‘fixed’ and we commonly refer to it as the ‘apex’. It’s not uncommon for drivers to focus on that to potentially ‘time’ the initial turn in and then focus on the exit soon after. There is little time for correction at this point anyway. Though the range of fixations very much differs between drivers.
We have covered some of the research of what has been observed in drivers and speculated on the purpose of various observed phenomena in a racing context. However, there has not yet been an extensive study into the exact eye strategies in elite racing drivers (the study referenced above cites the low amount of subjects to study), and more importantly discussion about why they happen in a scientific context. Further research needs to be done to observe whether drivers use multiple eye-strategies during the process of learning as familiarity with race vehicle and environment grows. We have already observed that low-skilled drivers use Tangent Point Fixation which appears to evolve into RTPF as skill increases.
Eye trackers have observed highly skilled drivers in action, but extensive analysis has yet to be performed especially with reference to relevant brain regions and functions. Whether the strategies themselves are required to be deliberately practised or whether they are a consequence of other functions in the elite driver’s brain is also not yet fully understood and will be discussed in another review.
1 – H Hicheur et al. 2005: Head Motion in Humans Alternating Between Straight and Curved Walking Path: Combination of Stabilizing and Anticipatory Orienting Mechanisms
2 – Land, Tatler. 2001: Steering with the head. the visual strategy of a racing driver.https://www.ncbi.nlm.nih.gov/pubmed/11516955
3 – Dr. Edmund Donges. 1978: A Two-Level Model of Driver Steering Behavior https://journals.sagepub.com/doi/abs/10.1177/001872087802000607
4- Peter M. van Leeuwen, Stefan de Groot, Riender Happee, Joost C. F. de Winter. 2017: Differences between racing and non-racing
drivers: A simulator study using eye-tracking – https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0186871&type=printable
Author: Alan Dove – GTS RS Driver Performance Consultant – firstname.lastname@example.org – 07549994245