Improvement of an Unsignalized Intersection Through Signal Optimization and Geometric Enhancement: A Case Study of Milan Chowk, Butwal

Abstract

In fast-growing Nepali urban centres, unsignalized junctions frequently struggle to handle mixed traffic volumes. Many of these locations are controlled only informally by traffic police, which produces irregular operations, high delays, and lengthy queues. Milan Chowk, a four-arm junction on the Mahendra Highway (NH-01) in Butwal, Nepal, experiences repeated peak-hour congestion and therefore offers a useful setting for testing systematic improvements. The present work examines the current operation of Milan Chowk and evaluates remedial measures with SIDRA Intersection software. Classified volume counts were gathered across a continuous 72-hour window; morning and evening peak periods were then extracted from the 15-minute summaries for detailed modelling. Existing police-regulated conditions formed the baseline. Two upgrade options were tested: Alternative I (cycle-length optimisation only) and Alternative II (added approach lanes on the north and south arms together with optimised cycle lengths). Performance was judged by average control delay, level of service (LOS), and 95th-percentile queue length. Cycle-length optimisation by itself (Alternative I) failed to relieve congestion adequately, especially in the evening peak when the junction remained at LOS F. The combined scheme of Alternative II performed markedly better. Cycle times of 110 s (morning) and 130 s (evening), paired with the extra lanes, cut average delay by 21 % in the morning and 45 % in the evening relative to Alternative I and reduced the 95th-percentile queue by as much as 56 %. The findings indicate that, at a high-demand unsignalized junction such as Milan Chowk, geometric capacity expansion must accompany carefully optimised signal timing if an acceptable level of service is to be reached; signal timing alone is insufficient.