Project: POWERED TWO-WHEELER INTEGRATED SAFETY (PISA)
Last update: 06/04/2010 09:50:52
The project supports the Departments objective of reducing the number and severity of road accident casualties through improved vehicle (Primary) safety. Transport Technology and Standards (TTS) has a broad based research programme and this project will positively contribute to safety issues which relate to motorcycling within the UK.
This project is also relevant to the objectives of the European Community's Sustainable Surface Transport Priority (SSTP) specifically objective 4 - Increasing road, rail and waterborne safety and avoiding traffic congestion.
A key part of this process is the development and use of new technologies to provide integrated safety systems for a range of Powered Two Wheeler (PTW) vehicles. These systems, aimed at accident avoidance and integration with secondary safety devices for injury mitigation, will need to be demonstrated and have their feasibility and effectiveness established through rigorous trials.
Crowthorne House, Nine Mile Ride, Wokingham, Berkshire, RG40 3GA
+44 (0)1344 773131
Cost to the Department: £212,110.00
Actual start date: 01 September 2006
Actual completion date: 23 March 2010
Powered Two Wheeler Integrated Safety (PISa) Final Report
Author: M McCarthy, W Hulshof and T Robinson
Publication date: 01/03/2010
Summary of results
- The principal objective of the PISa project was to identify and develop an effective Powered Two Wheelers (PTW) integrated safety system which includes active system technologies. The system developed was to be focussed on vehicle technology, rather than infrastructure or personal protective equipment solutions. The integrated system developed in the project was aimed at improving primary safety, but also allowing links to secondary safety systems, such that injury severity could be mitigated in those accidents that the technology could not completely avoid.
Accident data from the APROSYS project was used to provide the information from a European perspective. However, this data source was insufficiently detailed to identify accident causation factors, and to assess the relative priority of PTW safety system countermeasures. Consequently, three other accident databases were used to access accident cases with in-depth information, representing accidents that occurred within the most important PTW accident types defined by the APROSYS project. The databases used were the UK On-The-Spot (OTS) database, the COST 327 database held by the Ludwig Maximilian University of Munich (LMU) and LMU‟s ongoing forensic database. As well as contributing and analysing OTS cases, TRL reviewed a further 70 accident cases from the UK Fatals database and developed a database to record details of the pre-crash accident phase.
Analysis of these cases reiterated the findings of previous research that the types of PTWs involved in accidents in Great Britain are biased towards larger capacity machines and are markedly different to that seen in the European sample. This is especially true when compared with southern European countries, such as Spain and Italy where mopeds are more prevalent in the population. In deriving the priority of potential safety systems, weightings representing the distribution of accidents within the APROSYS accident types in both Europe (APROSYS) and Great Britain (Stats19) were used.
A sample of in-depth accident data were taken from the three databases and the accidents reconstructed to determine the accident mechanisms. Objectively defined system characteristics were then combined with engineering judgement to assess the types of active safety system that might avoid or mitigate accident outcome. A list of 43 functional requirements developed by the PISa consortium was matched to active systems, and these were assessed for each in-depth accident case reviewed. A series of inter-team workshops were used to co-ordinate and validate the active system characteristics, the accident reconstruction, and the overall assessments made by the participants of the consortium.
The assessment of the potential countermeasures was not limited to the capability within the PISa consortium, although the final selection of systems was inherently limited by this factor. Priority lists for the PISa project were developed based on the OTS and COST 327 in-depth cases. The top five priorities (using Stats19 weighting to APROSYS accident scenarios) were:
Warn other vehicle of PTW presence - PTW emit signal/other vehicle receive signal or other vehicle to detect PTW;
Automatically stop other vehicle without input from driver - Stop other vehicle (autonomous braking);
Communicate and warn PTW that vehicle travelling from left, right or oncoming is crossing PTWs path - Other vehicle emit signal/PTW receive signal;
Automatically stop PTW without input from rider - Stop PTW (autonomous braking);
Detect and warn PTW that vehicle travelling from left, right or oncoming is crossing PTWs path - PTW to detect other vehicle and warn rider.
A questionnaire was also developed to survey PTW riders and gather information on riding behaviour, experience, risk awareness and acceptance of a range of safety systems on PTWs.
TRL analysed fatal accidents from Great Britain to supplement the other in-depth cases (which were, in general, lower severity accidents). The prioritisation of the potential countermeasure systems from this analysis was different to that found by the OTS ranking. Fatal accidents tended to be higher velocity impacts, which made these more difficult to avoid. The countermeasures identified were, therefore, more focused on mitigation (secondary safety) or tertiary safety (post crash) solutions. It should be noted that eCall was the highest ranked system for fatal accidents. However, this benefit was considered likely to have been overestimated because the response time of the emergency services or whether the casualty died at the scene was unknown. The assessment of eCall was made solely on assumptions relating to the accident circumstances and that more rapid medical treatment has the potential to improve injury outcome. The proportion of these fatal accidents which might realise an improvement in injury outcome was not assessed, and may be lower than the "target population‟ estimate.
While the list of 43 functions was prioritised and the selection of functions made primarily on this basis, other factors influenced the final selection. Those functions that were out of scope of either the PISa project remit or the capabilities of the PISa project partners were removed from the prioritised 43 functions. The in-depth priorities and the fatals priorities were then considered concurrently and safety systems to be developed by the PISa project were chosen with consideration of both sources.
There were some functions that ranked highly in the fatal priority list that were not considered by the project. These include:
Restrict PTW to posted speed limit - ISA speed restriction;
Help rider to maintain steering and prevent loss of control;
Advise rider of approaching permanent hazard (sharp bend, steep decline) - GPS receiver on PTW.
These functions were excluded because they ranked very low down the in-depth priority list. However, because they are high priority systems for UK fatal accidents, they are particularly relevant to the UK situation, even if they have not been chosen for inclusion in the PISa integrated safety system.
The systems that were recommended to be taken forward by the PISa consortium were:
Automatically stop PTW without input from rider - Stop PTW (autonomous braking);
Detect and warn PTW that vehicle travelling from left, right or oncoming is crossing PTWs path - PTW to detect other vehicle and warn rider;
Avoid locking of wheels - ABS;
Balance front to rear braking force - CBS (combined braking system/linked brakes);
Reduce closing speed - ACC (adaptive cruise control);
Amplify braking force - Brake Assist, EBS (enhanced braking system);
Improve PTW conspicuity - Special fairings/PTW conspicuity.
An initial benefit assessment was carried out to provide an estimate of the magnitude of the casualty saving attributable to each potential safety system, taking into account accepted casualty valuations. This confirmed the selection of the systems by the PISa consortium and highlighted that the seven systems selected have significant target population benefits.
With the exception of ABS (which is an established technology and was therefore not included) and a system to improve the conspicuity of the PTW (which relies on a reaction of the other road user), the recommended system types were implemented by work packages 4 and 5 onto two different Powered Two-wheelers (PTWs): a 500cc Malaguti Spidermax and a 160cc TVS Apache. The TVS was equipped with CB and Semi-active suspension, with all other systems fitted to the Malaguti. The systems included in the PISa integrated PTW safety systems were as follows:
. Active braking (AB) - Low-level autonomous braking at 0.25g with pre-warning
. Enhanced braking (EB) - Additional braking effort applied (brake assist) in an emergency
. Combined braking (CB) - Distribution of brake forces between the front and rear wheels
. Distance Support (DS) - Automatic throttle inhibition for critical time headways.
In addition, a further system was developed which was not specifically recommended, but whose action complements the behaviour of the PTW under braking
. Active and semi-active suspension - adjusting suspension characteristics to complement heavy braking performance
These systems were evaluated against a defined test plan to measure the performance of the systems. The aim of this was two-fold; firstly to allow the performance of the systems to be used in the final benefit evaluation which aims to inform on the likely casualty reduction relating to system fitment and the costs of the systems, and secondly, to allow improvements to be recommended to the final prototypes.
The systems fitted to the Malaguti exhibited technical problems and the overall system could be reasonably described as being still in the development phase. The selection of valid tests (those in which the system was functioning) meant that significant quantities of test data were excluded from the analysis. Therefore, the analysis presented reports on the "best case" performance of the systems and excludes all incidences were there was an identifiable system failure.
Results showed that for Active Braking (AB), the mean triggering reliability decreased with test speed, from 91% at 35km/h to 58% at 55km/h. Reliability increased with increasing AB trigger setting (range 58% to 89%). This showed that the system was more reliable when set to activate very close to impact (mitigation system). At settings consistent with accident avoidance, the reliability was low. For tests in which the system functioned correctly, the distance at which the AB system triggered was between 23.3m and 6.2m prior to impact, depending on the trigger setting, and between 11.2m (for tests at 35km/h) and 33.4m (for tests at 55km/h) for AB trigger setting 3.
For the tests involving crossing, the "crossing scenario", the reliability of the system was low for the lowest (avoidance) AB trigger settings, with the system triggering on the test object in 9% of tests. At full mitigation (AB setting "9"), the reliability was 91%. For these tests the AB triggering distance was between 20.67m (Time to Collision, TTC =1.16s) and 5.48m (TTC=0.44s) depending on the AB trigger setting. It is recommended that the decision logic is improved so that the performance in this important accident type is improved.
Test riders preferred AB trigger setting 5 (activation of AB when approximately 0.5g required to avoid impact). Test riders were content with the level (c0.25g) of the AB system.
For the tests on the Enhanced Braking (EB) system, results showed that, on average, and with EB on, a rear brake application by the rider delivered a stopping distance between 2.1% and 6.8% better than a test rider applying both front and rear brakes. This shows that the PISa EB system is very effective at providing emergency braking. If such a system was introduced onto a production bike it would most likely be in conjunction with ABS and would mean that the even at high braking forces, the wheel would be prevented from locking.
The mean results also showed a weak trend for better braking improvements with increased suspension settings, although statistical testing showed no significant differences by suspension setting for the "EB on" group. The results also show that for the EB off group, the test rider (an experienced rider performing a planned braking manoeuvre) was able to improve stopping distances on average by between 15.1% and 19.4%. For this group, the suspension at medium or maximum, had a significant (P<0.001) effect on stopping distance. This showed that when the braking response of the rider was to use the front and rear brakes, the addition of active suspension makes a significant difference to the overall stopping distance; this is explained by the suspension reducing dive and giving the rider more confidence in applying the brakes, essentially realising more of the PTW‟s braking potential.
For the Combined Braking (CB) system, results showed that the addition on CB had a significant (P<0.001) positive effect on stopping distances, with average distances decreasing by 18.5%. The addition of semi-active suspension did not have any significant effect on stopping distance, although a weak trend for a benefit was noted in the average stopping distances with the suspension.
For the Distance Support (DS) System), results showed that there was no clear improvement of the car-following task. There were two tests performed, one with a prescribed car-following distance and one with a car-following distance of which the rider felt comfortable with. There was no improvement of the car-following performance shown in the first test. The results of the second test showed that the rider was more capable of following the lead vehicle with the DS system compared to without the DS system. The track tests showed that the results were influenced by false alarms, with these being caused by falsely detected objects seen by the laser scanner. According to the subjective measures, the DS system appears promising for improving PTW safety. However, the system requires further improvements to reduce the false alarm rate.
Work package 6 also made an estimate of the casualty benefits of the PISa systems. An evaluation methodology was applied, and predictive estimates were made for the target populations applicable to each safety system. Further estimates using a speculative range of system effectiveness values were used to examine potential break even costs for the individual PISa systems.
The benefit estimate methodology was limited by the information available; the tests planned within the PISa project, focussed on system validation with consideration of system evaluation. Although these tests showed that the system had the potential to function as intended, and demonstrated good performance improvements in specific, "uncluttered" test conditions, the test results could not be related to real-world system effectiveness. This was because insufficient information exists with which to robustly define the system effectiveness at the level required to perform a full cost benefit analysis. This was for a variety of reasons:
. Information relating to the chronology of accident events are missing from the retrospective accident data, meaning that the magnitude of performance increase provided by the system could not be objectively defined.
. Information on how the systems perform on other, more representative PTWs, was unknown.
. The real-world effectiveness is influenced by the rider. The PISa testing used test riders who were informed about the types of test being conducted. Therefore, no data was available regarding how riders of different ages, levels of experience, gender or attitude react to the PISa systems in "on-the road" conditions, or whether the systems have different levels of effectiveness for different users.
. For those systems were the performance can be directly compared to what the rider did in an accident situation (for example the braking performance for CB), while the results suggest that the impact speed can be reduced by approximately 18.5%, the effect on injury outcome cannot be objectively estimated. This is because the impact object and rider kinematics and trajectory are unpredictable and no injury risk functions exist with which to quantify the effect on injury level of reducing the impact speed, although the overall effect of reducing the impact energy is of course theoretically beneficial.
For these reasons it proved very difficult to extrapolate the track performance to a larger accident population. This, coupled with the fact that the system was focussed on accident mitigation (and not avoidance) made estimation of the injury benefits difficult. However the EU-27 target populations (numbers of PTW casualties who could be influenced assuming 100% system effectiveness) for the individual PISa systems were estimated. Corresponding annual GB values were estimated as:
a. Distance Support System: Fatal 7; Serious 82; Slight 254
b. Active Braking (following): Fatal 1-9; Serious 12-117; Slight 23-323
c. Active Braking/Warning (crossing): Fatal 2-38; Serious 9-464; Slight 11-1,267
d. Enhanced and Combined Braking: Fatal 8-72; Serious 355-76; Slight 63-1,890
A sensitivity analysis on the system effectiveness was carried out to enable an indicative, "what if" analysis, covering subjective estimates of system effectives between 10% and 50% of the target population. For these casualties it was assumed that the casualty severity outcome was reduced by one level - i.e. from fatal to serious, serious to slight and from slight to no injury. This analysis yielded the following Great Britian annual casualty benefits:
a. Distance Support System: £20 million (low effectiveness); £102 million (high effectiveness)
b. Active Braking (following): £2 million - £22 million (low effectiveness); £10 million - £112 million (high effectiveness)
c. Active Braking/Warning (crossing): £1 million - £93 million (low effectiveness); £5 million - £466 million (high effectiveness)
d. Enhanced and Combined Braking: £5 million - £151 million (low effectiveness); £26 million - £754 million (high effectiveness)
This speculative analysis indicates that the break-even costs for the system are relatively low and are likely to mean that scanning and autonomous decision making actions are not cost-beneficial without reductions in cost or specific studies or field-operational tests showing that the real world effectiveness is greater than has been assumed.
It should also be noted that the systems were necessarily validated in the test phase as independent systems and the analysis reported here reflects this. However, when implemented, the system is an integrated system and the benefits predicted are therefore not cumulative; the target populations for the individual system types will in practice have some overlap, although without further data the extent is difficult to predict.
The PISa integrated system has been developed successfully and has shown demonstrable performance level increases in specific test conditions. PISa has shown that the application of safety technologies to assist the rider has been shown to be an effective strategy to mitigate the severity of critical situations and the PISa prototype shows potential for reducing European PTW casualties. Although good performance in test conditions is considered very likely to translate into real-world casualty savings, the estimates for the European population cover a wide range and are limited by the extent and quality of the effectiveness information available.