Summary This report describes research work conducted on the audibility and sound localization of two main types of backup alarms (tonal [beep...beep…beep signal] and broadband [shh…shh…shh signal]), the performance of alarms with automatically adjusting levels and the effect of the alarm mounting location on sound propagation behind vehicles, ultimately to promote optimal use of these devices and minimize the risk of accidents related to vehicles operating in reverse. The first two parts of the study required the participation of human subjects in the laboratory, while the last two required noise level measurements using alarms and vehicles in both the laboratory and the field. Several findings were derived from the various trials. Both types of alarms remain audible at levels well below the background noise level and elicit responses at levels slightly below that of the background noise. Moreover, the time it takes to move out of the danger zone is relatively independent of the type of alarm. Regarding localization of the sound emitted by a single alarm, broadband alarms were found to be markedly superior to tonal alarms; regarding localization of the sound emitted by two alarms triggered simultaneously or nearly simultaneously, broadband alarms were again found to offer a clear advantage. With regard to alarm level adjustments of self-adjusting alarms, they take place from one cycle to another, based on measurements of the background noise and alarm signal taken immediately before each cycle. This signal-to-noise ratio (SNR) can vary significantly over very short distances. The alarm mounting location on vehicles also raises concerns. The results of this research for five different alarm mounting locations show that sound propagation behind a vehicle is much more uniform when it comes from a broadband alarm. The best way to prevent accidents caused by reversing vehicles remains that of minimizing both the use of backup manoeuvers, and, wherever possible, the presence of workers or passersby in the vicinity of vehicles that have to back up, by implementing efficient traffic circulation plans. Even so, backup alarms are still a means used to alert nearby individuals of a danger. To ensure safer (or at least optimal) use of backup alarms, a number of recommendations were made: (1) alarms should be placed in an optimal mounting location, i.e., behind the vehicle, directly visible by workers in the danger zone, and at a height of 1 to 2 m; (2) ISO Standard 9533 should be used to guide sound level adjustments for backup alarms, including all sources of noise around the vehicle when measuring the background noise; (3) a maximum backup speed of 12.6 km/h should be respected to ensure the correct SNR ratio prescribed in ISO Standard 9533 and ensure a minimum reaction time of 2 seconds (compliant with SAE J1741); and (4) the use of broadband alarms should be favoured in situations where several vehicles are performing backup manoeuvers simultaneously and to minimize the risks associated with erroneous sound localization. Lastly, the recommendations put forward in this report are based on results obtained with people who have normal hearing (subjects with no hearing loss) and subjects not wearing hearing protection devices or safety helmets (hard hats). The effect of these conditions on the perception and sound localization of backup alarms has yet to be investigated.