Towards understanding and improving the safety of informal public transport in Sub-Saharan Africa : a real-time sensing and reporting system for minibus taxis

[摘要] ENGLISH ABSTRACT: Developed countries have led the way in the implementation of Intelligent TransportSystems (ITS), with the main objectives of improving road safety and efficiency. CurrentITS solutions are heavily dependent on advanced and expensive technologies, and do notnecessarily meet the unique requirements of public transportation in Sub-Saharan Africa(SSA). The informal minibus taxi sector dominates public transport in SSA in general,and in South Africa in particular, and is notoriously dangerous – leading to many fatalitiesannually.This work presents the design and testing of a sensing and reporting system for publictransport in SSA. The system contributes to improving the safety and efficiency of minibustaxis in SSA. The system provides three core functions, namely, reckless driving detection,multiple occupancy detection, and wireless reporting to a visualised online platform.The reckless driving detection system implements a novel model that augments inertialvehicle acceleration data with GPS speed information. The model is based on standardsused in road design, and takes into account the relationship between a vehicle's tyres andthe road surface. A lateral acceleration threshold, which is speed dependant, and a longitudinalacceleration threshold are suggested to detect reckless driving. Acceleration datais filtered to remove both high-frequency noise and zero-frequency offset, and comparedto the thresholds to detect reckless driving events.The occupancy detection system detects multiple occupants in the minibus taxi usinglow-cost capacitive sensor electrodes, which utilise the electrical field properties ofa human body for presence detection. A simplified mathematical model was created tocalculate the expected capacitance on the occupancy sensor electrode, and was comparedto the measured capacitance in the minibus taxi. The theoretical and empirical resultsdemonstrate that the capacitance of an occupied seat is more than double that of anunoccupied seat. Occupants were clearly detected in various scenarios, such as differentoccupant sizes, water on seat, various seated positions on the sensor, etc.The wireless reporting inside the vehicle is implemented using a ZigBee network onan Arduino platform. The extra-vehicle wireless reporting uses the existing SSA cellularnetwork, and the online data visualisation is implemented on Trinity Telecoms' SMARTplatform. The complete sensing and reporting system is implemented as a prototype, andtested in South African and Ugandan minibus taxis.The results demonstrate that the system achieves the set goals, and could be used tomake transport in SSA safer and more efficient.

[发布机构] Stellenbosch University