Improving intersection safety through variable speed limits for connected vehicles

Brief Description of Research Project We seek to improve safety around intersections by using connected (partially or fully) autonomous vehicles (CAVs) to reduce variations in traffic speeds. Previous work has proposed variable speed limits (VSLs), which can be used to smooth speeds to improve safety. VSLs are used to reduce variances in speed which leads to fewer short headways and fewer incidents. For human drivers, VSLs require a variable message sign, which restricts the spatial locations of speed changes and is limited to human precision in speed (typically increments of 5mph). VSL effectiveness is also limited by driver compliance. These limitations can be overcome by CAVs; VSLs can be transmitted using infrastructure-to-vehicle communications, and are therefore not limited to specific locations. We assume CAVs (or partial-automation such as Tesla autopilot) can be programmed to obey VSLs. In the short term, CAVs will comprise only a fraction of the vehicles on the road. However, CAVs obeying VSLs travel slower than the free flow speed, thus forming a moving bottleneck for vehicles behind them. Human-driven vehicles following a CAV are forced to comply with the VSL unless they can pass the moving bottleneck, even if they are not aware of the VSL itself. Consequently, a small fraction of CAVs obeying VSLs should cause a disproportionate impact on speed variations. We will create a VSL control policy and evaluate the safety impacts at various CAV market penetrations.