LIKE many countries, South Africa has been grappling with the challenge of reducing deadly and costly accidents where people cross railway infrastructure. According to the country’s Rail Safety Regulator (RSR), 453 people were killed on South Africa’s railways in 2015-16 including 370 people who were struck by trains. Level crossings contributed 80% of the risks in the external environment in terms of cost incurred by the operator and injuries or loss of life. RSR recorded 83 level crossing incidents in 2015-16 with five fatalities and 27 injuries.

SidBetween 2008 and 2014 Transnet Freight Rail incurred annual financial impacts averaging Rand 434m ($US 35.3m) a year as a result of collisions, derailments, level crossing accidents and people struck by trains. In response to these diverse safety challenges, the Mechatronics and Micro-Manufacturing division of the Council for Scientific and Industrial Research (CSIR) has developed an autonomous rail vehicle to reduce the risk of collisions and inspect infrastructure for defects that could ultimately lead to derailments.

The Survey Inspection Device (Sid) was initially conceived as a locomotive-mounted unit, but subsequently developed into an autonomous vehicle to minimise stopping distances. Sid is designed to run 1-2km ahead of a train, relaying images to the cab of the locomotive. If an obstacle is detected, Sid notifies the driver with live video feedback.

Sid can also warn motorists at level crossings of the approaching train. The device waits in the middle of the crossing until the train is in sight of the road vehicle, before accelerating quickly away. Visual recording systems onboard record all activity outside the device while high-intensity warning lights and a voice communication system warn road users to stop.

The aim of the project was to develop a low-cost, low-weight vehicle capable of operating at up to 60km/h with a stopping distance of less than 100m and a wireless control range of 2km. To minimise weight, Sid runs on an aluminium chassis, while fibreglass was chosen for the bodyshell as it can withstand the heat generated by a petrol engine and provides a strong basis for mounting sensors and lights. Durable polyurethane wheels were chosen instead of steel to further reduce the weight of the device. These can withstand temperatures ranging from -20oC to +200oC and the riding surface of each wheel can support a 136kg load.

An onboard computer runs control algorithms and is connected to the locomotive via a continuous long-distance long-bandwidth LoRa wireless communication link. The 868MHz ultra-high-frequency (UHF) link receives speed commands and sends back data on the status of the vehicle.

To detect objects on and around the right-of-way, Sid is equipped with a laser line scanner, as well as a multi-beam laser scanner for 3D mapping of structures and objects. A thermal camera is used to aid obstacle detection at night.

In inspection mode, the device continuously monitors the condition of the track, looking for defects while measuring the gauge. Sid can also be used to measure ballast profile and trackside structures. A combination of lasers and cameras are used to measure rail profile, enabling Sid to detect rail head defects such as corrugation and spin marks, while the CoroCam multispectral camera developed by CSIR is used to inspect overhead catenary. CSIR is also experimenting with ground-penetrating radar to monitor for defects in the subgrade beneath the ballast.

When a defect is found, the nature of the fault and its GPS location is stored and digitally communicated to a command centre.

Data gathered by Sid can be transmitted to the shore via a Wi-Fi link at the depot, GSM in the field, or a satellite link if all other signals are lost.

Testing of the prototype began in April 2016 and Sid was publicly unveiled at the Africa Rail conference and exhibition in Johannesburg two months later. CSIR says the vehicle has been continuously upgraded over the last two years. This has included migration from a petrol engine to a fully-electric propulsion system, and a major upgrade of the sensor data collection software.

Tests were carried out at Transnet Freight Rail’s Pyramid South facility near Pretoria, where Sid was tested at up to 60km/h. During these tests CSIR monitored the behaviour of the wheels, and tested software communication between the vehicle and the operator. Obstacles were then placed on and around the track to simulate people, animals and cars. These tests demonstrated that Sid was capable of detecting, identifying and classifying stationary and moving objects.

Currently Sid activates track circuits as it travels ahead of the train but CSIR says it is now discussing with RSR how Sid should interact with signalling systems. CSIR has also involved RSR in the design process for Sid to ensure the vehicle conforms with international safety standards.

CSIR says it is looking for investment to further develop the concept.