Written evidence from Neil Cullen (SRN 056)

Summary:

1. I concentrate on ‘self-help’ measures for individual drivers to reduce congestion.
2. I focus on the ‘Fundamental Diagram of Traffic Flow’1, using the stopping distances (‘thinking’ and ‘braking’ separately and combined) from the current UK Highway Code2.
3. Intelligent use of this diagram, including its possible incorporation by existing in-car technology, can reduce congestion and improve safety thus reducing delays and costs caused by collisions.
4. A key point is that on open roads drivers should aim to keep moving, ideally at 16 mph or more, rather than driving up too close to a tailback. Stop/go driving seriously reduces lane capacity.
5. This can be achieved by looking well ahead, slowing down early and keeping a flexible reservoir of space ahead. Some drivers sabotage this strategy and such behaviour should be discouraged.
6. ‘Tailgating’ in groups of more than a few cars carries a serious risk of collision at the rear of the group , because the ‘elastic band’ effect magnifies small changes in the speed of the lead vehicle.
7. The Committee can help reduce congestion and improve road efficiency by publicizing the effects and encouraging production of education and training material by public service broadcasters to make the basic phenomena more widely known and understood by drivers.
8. Consideration should be given to modifying the Highway Code and incorporate the congestion-busting measures I have outlined.
9. Communication and co-operation, as opposed to command, control and coercion, need greater emphasis. Effective road use requires all users to work together as a team.
10. Co-operation rather than competition in the use of public road space has the potential to maximize efficiency while keeping costs lower and reducing the need for new infrastructure.

1. My submission addresses Item 4a of the Committee’s Terms of Reference:

“How can we improve the reliability and efficiency of travel on the strategic road network?”

2. I am a private individual, semi-retired and currently a self-employed part-time DSA Approved Driving Instructor (Grade 5). However I have been a Chartered Professional Engineer. My degree covered Mathematics and Civil Engineering, including Operations Research and Fluid Mechanics, and I have been a member of the Institution of Civil Engineers, the Chartered Institution of Water and Environmental Management, and the Chartered Management Institute, and also worked as a Risk Management Practitioner in the Water, Railway and Power industries.

3. My experience of the strategic roads network is mainly as an inquisitive regular and frequent user. However I have been interested in road safety since schooldays. I have recently completed a monthly series of magazine articles3 in a popular style, to assist driving instructors in explaining the principles of risk management and application of the laws of physics to driving safety.

4. I have followed the history and development of the M25 motorway since its inception right up to the installation of variable speed traffic management, and I have also experienced the systems now running on the M42, and observed the recent installation of overhead gantry control systems on the M62 near Leeds where I now live. As a user I generally endorse their effectiveness, though there seem to be teething problems until users learn to appreciate their value, and work with them.

5. However I have observed that in the UK we seem to be reaching the limit of what can be achieved by ‘hard’ engineering and believe that the application of ‘soft’ methods, involving different technology and better driver understanding, could be significantly more cost-effective. Strategies of ‘control, command and coerce’ need to be complemented by ‘communicate and co-operate’.

6. I aim to present clearly the principle which underlies many of the advanced systems in place to control UK traffic. I believe that appreciating the basic numbers buried in the theory will provide drivers with self-help tools to enable them to reduce delay and congestion for themselves, not only in situations where control technology is not provided, but also on managed networks where better understanding will enable drivers to work ‘with’ the system and its operators.

7. It is obvious that many regular car drivers already understand the principles and apply them intuitively while driving. Failure to appreciate one particular aspect, namely that there is an ‘optimum’ speed for heavily congested traffic, is the primary cause of ‘invisible’ traffic blockages.

8. There is a simple antidote, but it requires driver co-operation. I hope that by publicizing the principles and encouraging wider dissemination and possibly incorporating them into future guidance and/or legislation, the Committee can directly assist in reducing congestion and improving journey times and user satisfaction without major infrastructure investment.

9. My proposals if implemented have the potential to reduce the number and severity of traffic collisions and thereby reduce the national bill not only for death, injury and damage but also in delay to the many other road users held up by traffic incidents.

10. In addition, the principles could be incorporated within existing ‘SatNav’ in-car technology to provide drivers with guidance on the optimum speed for each leg of a journey in order to minimise overall congestion and optimize journey times, without taking ultimate control from the driver.

11. My starting point is the well-known Highway Code Rule 126.

12. The Highway Code numbers are based on two working assumptions: (a) the driver’s reaction time, from an incident occurring to when the brakes are applied, is 0.68 seconds and (b) the braking force is 0.67g, i.e. the grip between the tyres and the road allows a deceleration force of ⅔ of gravity.

13. The most important point is the ‘square law’, i.e. when the speed is doubled, the braking distance is four times greater; when the speed is trebled, the braking distance is nine-fold greater, and so on.

14. Safety advice is that drivers should maintain a gap between vehicles of the full stopping distance. It is of course possible to drive with a smaller gap, but in feasible cases when either the vehicle ahead sheds part of its load, or hits an obstruction and spins around, or stops ‘dead’, the full gap is safest. At 70mph this safety is achieved with a gap of 3 seconds (The Highway Code recommends at least 2).

15. The carrying capacity of a carriageway is given by a simple rule, identified in 1933 in the USA by Greenshields7. It states that Q=V÷S, i.e. the Quantity of vehicles per hour increases with the Velocity of the cars and decreases as the Separation gets larger. This creates temptation to drive too close.

16. If we plug the UK Highway Code equations in to calculate Greenshields’ S, the result is shown below. The green line [8] shows the carrying capacity using the full Highway Code stopping distance, with an allowance of just under 4m for the average length of the cars. The maximum carrying capacity occurs at a speed of 16 mph, and is slightly more than 2000 vehicles per hour (vph). 1800 vph is equivalent to a 2-second frequency. The blue line [9] shows the theoretical capacity if drivers adopt a strategy of relying on looking well ahead and anticipating problems in time to slow down early.

17. The red line [5], a highly risky but feasible strategy shown only for comparison, tells what happens with a 1-second gap regardless of vehicle speed. At (say)100mph the capacity is 3300 vph.

18. From the point of view of congestion and road efficiency, the crucial point is that the capacity graph is ‘bi-stable’, i.e. as high speed traffic slows down, the road capacity increases, but if drivers get so close that they have to slow below 16mph, the road capacity falls off very quickly, resulting in traffic coming to a complete stop – the ‘phantom traffic jam’ ably illustrated on YouTube4 by the BBC’s Andrew Marr with the aid of Bristol University researcher Eddie Wilson8.

19. Many experienced drivers try to adopt a strategy of “look well ahead and slow down early so that you don’t have to stop”. Unfortunately, other drivers destroy these efforts because they have been conditioned by ‘stop/go’ driving conditions in towns into driving up close and forming a queue.

20. The graph also illustrates what controllers are trying to achieve by ‘variable speed limit’ overhead gantry systems: they are mobilizing the elasticity in carriageway capacity by reducing traffic speeds, in order to accommodate extra demand.

21. Committee members with experience of driving on congested motorways will recognize in the graphs the intuitive effect which we all notice: as the road becomes more crowded we slow down because we do not feel safe at higher speed with smaller gaps. The road becomes ‘self-managing’ – unless a driver gets so close to the vehicle ahead that he/she feels the need to brake.

22. These findings lead me to ask two questions: (a) do we really need expensive control infrastructure to manage traffic, when we could use the alternative of better educating and training drivers? and (b) could we provide better information to drivers about the local level of congestion, so that they can anticipate likely conditions ahead, then look further ahead and plan their  speed?

23. The data above can be re-plotted on different graph axes to show what is often called the Fundamental Diagram of Traffic Flow1. An instance of this using UK Highway Code criteria is in Figure 3 below. This is a very powerful display.

24. At a glance it gives vehicle spacing and resulting lane capacity for any given forward speed.

25. Thus, at 50 mph with 20 vehicles per kilometre (= 2 cars per 100 metres shown on top axis) the road can carry 1500 vph using the ‘safe’ full Highway Code stopping distance – the green line.

26. With the simple addition of a proximity sensor to tell it how close I am to the cars ahead of and behind me, my ‘SatNav’ could easily compute a ‘congestion index’ telling me the state of the traffic. There are two possible obvious forms a congestion index could take, either the lane occupancy in thousands, or the percentage of theoretical capacity actually in use. I prefer the former.

27. The Committee wishes to ponder the use of new technology: if some in-vehicle SatNav systems equipped as suggested above were to feed back real-time information, then a permanently up-to-date country-wide data bank of information could be maintained and re-transmitted to drivers.

28. The Fundamental Diagram gives additional information relevant to road safety.

29. The top axis indicates traffic density in vehicles per 100 metres; thus 25 vehicles x 4 metres long represents – literally - ‘bumper to bumper’. If a stream of traffic maintaining a 1-second separation (the red line in Figure 2) brakes to a halt, stopping exactly bumper to bumper, then a ‘shock wave’ travels backward towards incoming traffic, as the queue of stationary vehicles grows.

30. This shock wave travels at 8.7 mph towards the incoming traffic. If drivers manage to stop with a greater safety margin, say 1 car’s length between vehicles, then the ‘shock wave’ speed is 17.5 mph.

31. Thus even a driver travelling at only 50 mph, following ‘too close’ but not fully alert, may discover a ‘shock wave’ of red brake lights coming towards him or her at a net combined speed of 67.5 mph.

32. Figure 3 also promotes understanding of the congestion which develops where traffic streams merge. A case in point might be the notorious northbound M5/M6 junction near Birmingham, filmed from a ‘cherry picker’ crane in the Andrew Marr film clip4 mentioned above.

33. The speed of traffic in the M5/M6 ‘merge’ zone is often around 30 mph. From Figure 3 we see that this corresponds to about 1800 vph. Suppose there are 1000 vph approaching at 70 mph on the M6, plus 800 vph approaching at (say) 60 mph on the M5. Simple addition confirms that, providing drivers observe a competent ‘merge’ technique, it should be possible to accommodate both streams without dropping below 30 mph. However, the blue curve shows that good driver skill and co-operation might theoretically facilitate up to 3300 vph with traffic speed reduced to 16 mph. [Note: this is a simplified calculation to illustrate the principle; in fact there are 3 M6 lanes and 2 M5 lanes at this point.]

34. I do not advocate such an approach to increasing road capacity, but it has long been evident that it is often difficult to ‘build’ our way out of traffic congestion. Accurate traffic forecasting is difficult, and effects such as generated traffic and increasing prosperity compound the problems.

35. Human ingenuity at the grass roots is a fertile source of cost-effective solutions, and Great Britain still has, in my view, a justified reputation in their production and profitable export of many. One example of ingenuity and adaptation can be seen every day on our motorways. Figure 4 above is a snapshot extracted from a measurement video I took from a motorway bridge on the M62 outside Leeds during 50mph speed restrictions during installation of new traffic management equipment.

36. The cars in lane 3 are travelling at 50 mph and have formed a cluster or ‘Peloton’5– a device used by successful cross-country cycling teams and ‘Formula 1’ aces to reduce wind resistance and gain an advantage. Off the race track, it does not gain an advantage in wind resistance, results in ‘tailgating’ which disconcerts most drivers, and is completely unnecessary :– note the huge amount of spare space behind and ahead of the peloton which could be used to spread out and reduce the risks.

37. The highlighted fourth car back is actually travelling dangerously outside the safety envelope defined by the Highway Code. It is easy to verify from Google Earth that the lighting columns are 50 metres apart, so it is plain that the local traffic density is four vehicles per 100 metres, as opposed to the recommended two. The calculated gap is 1⅛ seconds. If the entire lane were full of ‘tailgating’ traffic, it would be carrying 3200 vph, but the probability of collisions would be high.

38. I need scarcely mention that when a major traffic incident occurs, the police often report that traffic was ‘travelling too close and too fast for the conditions’ – an everyday story of motorway folk.

39. Given such driving behaviour, these major incidents are mercifully - and amazingly - few. However even a ‘minor’ 3-car shunt on a busy motorway causes major disruption and delay. Such shunts are in my view the result of ignorance of basic driving physics, and carelessness. They cost us all dear.

40. I believe that the Committee can and should contribute to the improvements in safety and traffic flow and efficiency, simply by publicizing the fundamental principles and encouraging further education and training of drivers who use our strategic roads network.

41. Possibly some further supportive academic and practical research is needed, but I believe that the subject is now ripe for some ‘public service broadcasting’ illustrating the problems and demonstrating the benefits and effectiveness of alternative driving strategies.

42. In a nutshell, the issue could be seen as the result of today’s drivers learning and being tested only in the stop/go environment of towns, with little appreciation of the different practices needed on high-capacity fast roads. The 16mph free-flow / congestion ‘tipping point’ is largely unknown.

43. As any team racing cyclist knows, the ‘tail end Charlie’ is at risk because of the ‘elastic band effect’5,8. The lead cyclists are in control, but any slight change in speed is transmitted and magnified back down the line, so that those at the back find it very difficult to react in time to avoid collision.

44. It can be demonstrated mathematically that even with a group of ten cars, drivers observing Highway Code guidance and travelling at 70 mph, if the lead driver reduces speed by slightly releasing the accelerator without applying the brakes, the tenth car in the peloton will almost inevitably collide with the ninth.  The effect can be observed. This risk should be publicized.

45. Another common practice, less dangerous but annoying and frustrating, is when drivers attempt good practice approaching a tailback, by matching the average speed of a slow-moving queue up ahead. They leave a ‘buffer gap’ (which grows and shrinks as the traffic ahead goes and stops). This strategy benefits all the users in the ‘informal convoy’ behind, because it reduces speed changes and increases fuel economy, also reducing the driver and vehicle stress of continual stop/go queueing.

46. What ruins the strategy for all the other road users is when a driver from another lane goes past the controlling ‘lead’ car then moves across, stealing the buffer gap to gain short term ‘advantage’. This overtaker then has to brake to a halt at the tailback, and causes a blockage, as previously described.

47. I am probably not alone in feeling that such selfish behaviour should be penalized. It is illegal to overtake the lead vehicle of a convoy through road works; perhaps this principle could be extended.

48. The skills to approach a queue gradually and avoid stopping if possible are not difficult to acquire, and can be learned in ‘town’ driving e.g. on the approach to traffic lights and/or roundabouts. Figure 5 illustrates the shockwave generated by traffic braking heavily to a halt; the tailback is growing at 4 metres per second (= 8.95 mph). Highway Code stopping distances can be read from the graph for Car 1: e.g. at 70 mph the braking phase covers 75 metres and takes 4¾ seconds.

49.                                        Risk Compensation (the ‘Peltzman Effect’6 ) is often evident. I therefore conclude with a freshly-prepared update to the UK ‘Fundamental Diagram’ to allow for the fact that in dry conditions on a good road surface it is possible to achieve a 0.9 g emergency braking force, as compared with the 0.67 g assumed in the current Highway Code. A recent talk to Leeds driving instructors by a local police collision investigator confirmed that the 0.9 g value is now in common practical use.

50. Some points to note on Figure 6 are that (a) the 0.9g ‘safe’ curve (dashed green line [10a]) indicates a possible 10% increase in maximum lane capacity to 2200 vph in clear, dry conditions, (b) the ‘risky’ curve (thin purple line [10]), which relies on excellent driver alertness and anticipation, gives an increase in maximum capacity from 3300 vph to 3800, an increase of 15%, and (c) the 16 mph ‘tipping point’ speed increases to 20 mph.

January 2013

References

1. ‘Wikipedia’ ‘Fundamental Diagram of Traffic Flow’
2. Highway Code, Revised Edition 2007, Rule 126
3. ADI News, Brighton BN1 45T, monthly magazine articles by N Cullen, August 2012 – May 2013
4. ‘Phantom Traffic Jam’ http://www.youtube.com/watch?v=goVjVVaLe10
5. ‘Wikipedia’ ‘Peloton’
6. ‘Wikipedia’ ‘Peltzman Effect’
7. Greenshields B D, Proc.  13th annual meeting of the US Highway Research Board, Dec 1933
8. Traffic Wave demonstration http://www.youtube.com/watch?v=19S3OdK6710