Abstract
The three important parameters describing the operation of a highway facility are speed, volume and density of the traffic stream. This report is concerned with the volume parameter. An efficient surveillance and control system for an urban freeway should have the ability to obtain information on vehicles at several locations. To locate detection equipment in each lane a several locations is anexpensive proposal. This paper describes the development of linear multiple regression models to predict total freeway volumes as well as individual lane volumes. Classification of vehicles along with the interlane volume is also considered.