Capacity and Behaviour on One-way Cycle Tracks of Different Widths

Abstract

The main objective of this study is to examine how widths of cycle tracks influence the behaviour, flow and capacity of bicycle traffic. Empirical data has been collected by video observations at 8 different cycle tracks of varying widths (1.85m-2.85m excl. kerb between cycle track and carriageway). The locations are characterised by high bicycle traffic volumes on the track, no traffic lights / junctions / bus stops / zebra crossing nearby, and no dividing verge between cycle track and carriageway. Speed and lateral positions has been measured for 8,925 cyclists. The average speed is 21.6km/h but differs slightly between locations. Women ride 2-3km/h slower than men but also with a smaller dispersion. The traffic volume does not affect average speed, but dispersion decrease with increasing bicycle traffic volumes. At narrow cycle tracks cyclists are riding closer to the footpath and closer to each other during overtaking compared to cycle tracks of a larger width. Car parking in the road side next to the cycle track reduces “the effective width” of cycle track with about 10-15cm. High bicycle traffic volumes are only observed in short time spans and it seems like the capacity limit is not reached. Flows as high as 20 bicycles per 10 sec are observed at a 2-lane cycle track (width: 2.35m) and still with an average speed of about 21km/h. Controlling factors are used when calculating capacity from short time spans, and the hourly capacity of a 2-lane cycle track has been estimated to about 3,000bicycles/h. The width does not affect the capacity much unless the number of lanes are reduced or increased. A cargo bike has an average speed of 16.3km/h and its headway is 1.3 times as big as the headway of a traditional bicycle. Due to the speed and the size of a cargo bike it reduces capacity equally to 3-4 traditional bicycles. Based on the data minimum and recommended widths of bicycle tracks are found.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 30 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.