The potential benefits of using resonant cycle times in traffic signal control on an arterial are investigated. Resonant cycles are cycle lengths that result in good arterial progression over a range of traffic flows. The notion of resonant cycle times contrasts with the prevalent adaptive control prac- tice of setting the arterial cycle length in proportion to flow levels at the most congested intersection on the arterial. This research was motivated by the development of appropriate adaptive algorithms for closed-loop system control in the FHWA ACS-Lite project. Simulation experiments with TRANSYT-7F for a four-intersection arterial and additional time– space diagrams demonstrate the characteristics of resonant cycle times and the substantial performance benefits that they may offer. In addition, a systematic method was developed to identify appropriate resonant cycle times and fine-tune a schedule for a time-of-day signal timing strategy. Selecting cycle lengths on a systemwide basis has a significant impact on performance. However, there is little agreement in the literature on the appropriate methodology for selecting either the system cycle or the number of subsystems into which a network of signals should be partitioned. In considering possible cycle time adjustment strategies for FHWA’s ACS Lite adaptive control software (1), the following two approaches are of interest: • Use critical intersection analysis, in which one calculates a cycle individually for each intersection of the arterial by using equations such as those of Webster (2) or the Highway Capacity Manual (HCM) (3). Then the longest cycle length of all intersections in the system is applied as the common cycle length for the arterial. • Use software such as TRANSYT-7F (4) or Synchro (5) for offline, systemwide optimization of timing parameters (cycle, off- sets, and splits) based on specified performance objectives, such as minimization of vehicle delay and arterial travel time. Although signal-timing optimization software such as TRANSYT- 7F and Synchro are commonly used as benchmarks for good arte- rial signal timing, a number of variations on the critical intersection analysis heuristic are still in practice. It has been noted that “it is simply amazing how often the cycle length is poorly set for system purposes” (6 ). In addition, most currently deployed multi-intersection adaptive control systems also utilize a critical intersection approach to online cycle time adjustment. In this study, the focus was on a popular cycle selection strategy, referred to here as the 90% rule. In addition, the cycle lengths generated by this strategy are referred to as 90% cycles. This heuristic is based on monitoring the degree of saturation for all intersections in the system (3). If any approach to an intersection becomes more than 90% saturated (and reallocation of splits cannot reduce this value), the common cycle length of the whole system is increased by a few seconds. Conversely, if all approaches are less than 90% saturated, the network cycle length is incrementally reduced. This cycle time adjustment strategy is used by SCOOT (7 ) and VFC-OPAC (8). Similar strategies, based on the degree of sat- uration or volume-to-capacity (V/C) ratio measures, are also used by SCATS (9) and the Los Angeles Department of Transportation’s adaptive traffic control system (10). In this study the characteristics of a system-friendly cycle length and its performance relative to strictly flow proportionate approaches such as the 90% rule are investigated. In particular, the notion of a resonant cycle is, simply put, a cycle length that accommodates good two-way arterial progression—an essential qualitative characteristic of well-timed traffic signals when feasible. The resonance (and non- resonance) of a cycle time stems from the principle that intersection spacing precludes good progression at certain cycle lengths, and conversely some cycle lengths necessarily preclude progression on an arterial because of their incompatibility with the network geometry. The second essential element of a resonant cycle is that it continue to provide good two-way arterial progression over a range of volumes. Although the cycle length may not be optimal, in a strict global sense, leaving the cycle length at this value may outperform typical heuris- tics used to modify cycle length in the context of adaptive control. The following study compiles some quantitative evidence to substantiate benefits of resonant cycles. RESONANT CYCLE STUDY This study of the potential benefits of resonant cycles is organized into three subsections: (a) the experimental design, (b) a look at the traffic flow mechanics that give rise to resonant cycles, and (c) the performance benefits quantified in terms of delay, travel time, and stops. Resonant Cycle Experimental Design Two versions of a simple four-intersection arterial were modeled in TRANSYT-7F, a signal-timing optimization and macroscopic simulation software tool: • A two-way arterial model featuring balanced traffic flow in both directions of the main street and • A one-way arterial model using two through lanes on the main street and a single through lane in each direction on the cross streets. Intersections were spaced 600 ft apart, and the same traffic flow was used in each lane. This model creates the effect of balanced volumes (and congestion) for the main-street and cross-street phases and most Resonant Cycles in Traffic Signal Control Steven G. Shelby, Darcy M. Bullock, and Douglas Gettman S. G. Shelby and D. Gettman, Siemens ITS, 6375 East Tanque Verde, No. 170, Tucson, AZ 85715. D. M. Bullock, School of Civil Engineering, Purdue University, 550 Stadium Mall Drive, West Lafayette, IN 47905. 215 Transportation Research Record: Journal of the Transportation Research Board, No. 1925, Transportation Research Board of the National Academies, Washington, D.C., 2005, pp. 215–226.