Vehicular Interactions For Traffic Demand Estimation Under Mixed Traffic Flow

Abstract

The efficiency of the transportation network depends basically on the accuracy of the estimation of the basic design input i.e. the traffic demand on the network and the relative variations of this demand on time-scale. The demand being a stochastically variable input requires an accurate assesment of inherent flow variables and vehicular interactions. The mixed traffic flow which is available on Indian city roadsj both on major urban centres and small towns, is peculiarly different than that occurring on such or similar right-of-ways in developed countries. Unfortunately, the approaches developed in major industrialized and developed countries do not fit to the analysis of mixed traffic flow on Indian roads due to the following specific reasons: (i) The basic mechanics underlying the flow of animaldrawn, manual-drawn and other slow moving twowheeler and three-wheelers is very little or incompletely understood in isolation and much less in conjunction. (ii) As such the flow process becomes a complex agglomeration of unknowns and very few known variables. ii (iii) The mixed traffic flow analysis being done around the world or available in literature from past work, basically centres around the accommodation or inclusion of commercial and/ or twowheeler motorized vehicles alongwith the passenger cars. As such, the basic terminology of 'mixed-traffic-flow1 carries different meanings, in India and abroad. (iv) The speed-differential amongst slow-moving and fast-moving vehicles, occupying the same right-of-way is so high (sometimes 5 kmph or less to 50 kmph or more) that actual total flow has very complex interactions. Such aggre gated flow remains almost a 'non-deterministic' quantity to-date. (v) As the vehicular incompatibility in aggregated flows increase, a slight variation in composition, operating speeds or network geometry magnifies into large variations in flow conditions and consequent levels-of-service rapidly change. This is extremely difficult to explain through avail able or developed approaches which basically represent the flow behaviour of automobile or motorised traffic. From the above discussion, it can be concluded that the assessment of total aggregated mixed traffic flow is Ill extremely important for any rational planning of the transportation network. The assessment of the above parameter is relatively simple as well as adequately developed for homogeneous flows or only automobile traffic flows including passenger car and commerical vehicles which is also termed as 'mixed-traffic flow' in reference to developed nations. A traffic Engineer has a two-fold responsibilityj- (i) To fit traffic to the existing network, and (ii) to adjust network to the traffic. The network traffic capacity is essentially the available 'supply' which has a maximum value restraint. The 'demand' in the network is the dis crete movement of objects, the vehicles. Obviously, the maximum-demand and supply are to be analysed to provide an optimum network. The volume, therefore, which is 'effective demand' is taken as 'traffic demand' on the network and has been attempted to be understood through this research programme/ project; especially for the type of mixed traffic flow occurring on Indian Roads. For traffic flow system in India, it has been estimated that some six to thirteen types of incompatible types of vehicles occupy the same right-of-way and move in aggregate flows. The vehicle sizes vary from pedalbicycle to large hand-carts and tongas in slow moving categories; and compact and small scooters and motorcycles IV to large commercial vehicles including trailor and semi-trailor combinations in automobiles. The speed differential being, speeds as low as 5 kmph or less to 50 kmph or more on urban roads. The roadway character istics varying from narrow single lanes to multiplelanes; and single lane-two way to multi-lane one ways: including divided and undivided roads. As such the cluster of variables is too large to be analysed by any simpler intuitive approach. Hence, the logical and ultimate choice is a comprehensive study. As such this research project is an attempt to develop and establish a set of relationships to define, estimate and accurately predict the traffic demand on a typical network and to indirectly link it to the 'supply' or the capacity through volume/capacity ratio estimation and level-of-service operating at a given time under existing traffic conditions. There is a multitude of categories or types of the vehicles, the network geometry and their related attri butes as explained earlier. However, within the scope of the time and resources availability and obvious financial and man-power restraints of tedious traffic data, collection programmes; the scope of this research project is: (i) To analyse the vehicular characteristics of major vehicle,types predominantly available on Indian roads with special reference to free-flow speeds, operating or restrained speeds, inter-vehicular spacings, clearances and manoeuvring. (ii) The analysis is carried out on divided and undivided highways or roads with typical Indian urban traffic. However, studies have been done on selected routes or sites, for reasons explained earlier. (iii) Incorporating the vehicular traffic character istics, a series of versatile computer pro grammes have been developed and successfully executed on IBM 370/145 system in Fortran IV language. (iv) The work enables assessment of traffic demand for varied traffic situations with considerable accuracy. This is the basic as well as ultimate objective as stated earlier in the programme. (v) Acounter-check to the accuracy of the developed methodology is provided through an indirect built-in level-of-service concept incorporating volume-capacity ratio and operating speeds as variables. (vi) The developed mathematical model describes a simple approach to finding solution of the VI problem; but developed through a series of tedious trials over a long period of time. (vii) The work explicitly brings out the need to discard the use of fixed P.C.U. orPCE's for converting mixed traffic flows of Indian type to aggregated equivalent homogeneous flows in terms of passenger cars. This is illogical , erroneous and misleading, as the equivalent effect of a vehicle changes very appreciably under varying conditions of traffic composition and flow. In this programme two unique and hertofore untried approaches in India, have been incorporated. These ares- (i) Time-Lapse photographic study technique has been modified to suit the need for study of heavy vehicular traffic on divided multi-lane roads carrying mixed traffic. The approach used in the study is, however, a. deviation from conventional perpendicular to-road-horizontal camera photography. In this study, the picture of a. relatively larger portion of road from an elevated point, is taken in the forward direction of traffic flow. Frame-by-frame analysis is done by a specially designed apparatus-attachment and a developed projectioncum- measurement grid. vii (ii) On undivided city roads, a. specially designed programme of 'road-trap-length' and a suitably superimposed and 'pavement-marked-grid• has been employed successfully to obtain a. large and useful traffic-data, almost comparable to the accuracy of work of divided roads. The developed experimental programme is versatile enough to be used anywhere; wherever similar traffic conditions exist. While similar programmes tried in developed and advanced countries, notably U.K., have simple arrangement for micro-analysis of a given vehicle; in this programme overall and macro-analysis of the traffic stream had to be performed due to largely varying vehicular character istics in the mixed traffic flow. As outlined and explained in the experimental pro gramme, the analytical programme is also divided in seq uential steps and then synthesised to obtain logical and reasonable results of practical as well as theoretical importance. Summarised programme is: (i) An initial programme to understand the basic flow process underlying mixed traffic flow is undertaken and its behaviour is understood on a. macro-level. (ii) The work is divided into categories, i.e., the divided and undivided roads and the flow viii variables are analysed with special reference to space-headways, spacings, lateral clearances and speeds. The programme is analysed simul taneously for speed-interactions and flow. The basic data input is derived from the experi mental programme as listed earlier, i.e. timelapse photographic study and 'pavement-grid' measurements. (iiO Mathematical models synthesising the above results are derived and built into a computer analysis programme to obtain the effective 'traffic-demand' in the flow of each vehicle type. (iv) The concept of 'pavement-occupancy' and vehi cular-interactions is successfully exploited to obtain 'demand' of each category or type of vehicle in the flow. (v) The analysis is initially done in trial stage on IBM-1620 computer system and finally on IBM-370/145 system in Fortran IV. The results obtained were validated. (vi) The analysis programme is dynamic in its behaviour and elastic, responsive, and sensitive to operating speeds of each vehicle type, traffic-stream composition, and speed of standard vehicle type .i.e. the passenger car IX and density-index defined in a particular way. (vii) The concept of level-of-service is introduced and built into the programme to provide an additional validation check on mathematical analysis, which has provided excellent results. The results may be summarised as: (i) A clear and better insight into the flow variables of mixed traffic flow occuring on Indian roads is now available in terms of vehicular-interactions, spacing, space head ways and gaps, operating and free-flow speeds, vehicular composition of traffic stream and vehicular pavement occupancy. This was not available in any specific terms to date for Indian conditions. (ii) The real and basic design input for transporta tion network planning, i.e. the volume of traffic or'effective-traffic demand' can now easily be estimated and the estimation can be updated as the basic input variables undergo a change and as such provides a tool for eval uation and monitoring of transportation net works also. (iii) Quantitative prediction of volume-capacity ratios and qualitative assessment of the X level-of-service provides an excellent tool, for transportation planning for mixed traffic flow networks, which was here-to-far not available in a. single meaningful methodology. As outlined in the results, this programme can provide an effective guideline for field implementation. The specific utility can be summarised as: (i) For improvement of existing urban transportation networks, the programme provides one simple assessment of present demand on the network and its level-of-service. (ii) The programme also provides an excellent tool to accurately and precisely estimate the demand for new networks, guidelines ofor critical operating speeds and predict the level-ofservice of resulting traffic operation.