Most train accidents are caused by derailment, which in turn is mainly because of equipment failure of either the tracks or the wheels/axles. MINDLOGICX' s train safety system will prevent such accidents by continuous online monitoring of the condition status parameters like vibration, temperature, etc. of axles, which reflect the condition of the wheels and tracks. It is well established that the Fourier analysis of the vibration data is a powerful tool to monitor the condition of axles and tracks. The growth of IP based wireless networks and wide spread use of wireless and IP based communication devices have given birth to a new business opportunity for development of remote monitoring systems using these technologies. The project envisages development & marketing of embedded hardware & software, wireless net connectivity, intranet real-time database servers and intelligent alarm systems to the railways worldwide. The business model is based on direct sales to railways and OEM sales to railway equipment vendors (vendors of axles, wheel systems, wheel bearings, etc). A small, embedded system with the sensor electronics, a Digital Signal Processor (DSP) and a RF wireless module is installed in close proximity of the axle.

The parameter sensor is mounted directly on the axle. The Fourier analysis is done in the DSP and the processed data is transferred to the aggregation module using a RF module (wireless LAN, Blue Tooth, etc.). The aggregation module can interface to a multiple number of sensor modules. The aggregation modules transfers the data aggregated from the multiple sensor modules to the internet access modules mounted on poles for say every 50 meters of rail track, whenever the train is in the vicinity of the poles. The Internet access modules are connected to the Internet and transfer all data to the central server. Instead of the Internet access modules, the aggregation modules can also transfer data to the Internet server using GSM / GPRS infrastructure.

Statistical software & expert system on the central server (rule based or fuzzy logic or a hybrid) will produce daily/weekly/monthly condition status reports for the monitored devices. The condition status reports shall provide an indication and chance for possible failures, if any.

Wireless inter-connect technologies (wireless LAN, Bluetooth, etc.)

Statistical methods for predicting failures

Rule based methods for predicting failures

Fuzzy logic based methods for predicting failures

Comparison of the three methods

Validation methods for transmitting sequence data via GSM/GPRS network

Validation of measured data from sensors; and rejection of incorrect data

The future application areas where most of the developed components viz., hardware & software can be re-used are in the high hazard industries like process industry, steel industry, power plants and electrical power distribution.