Two sensors with different output polarizations (linear and circular) were installed on temporary light support system on the front of engine. Using of two sensors allow to detect / cut off wrong reflected signals from the land and metal parts of railways lines. After setting at supports, systems were pointed in direct way onto a target at 150m distance. Data processing, saving and visualization from two sensors were arranged on laptop is placed into cabinet of the engine. Parallel all tests were stored on digital camera. Below you can see pictures with installed sensors.

To avoid two-times reflected signals in one sensor we used clockwise circular polarised wave for transmitting and anti-clockwise polarised wave for receiving. One time reflected wave will be received, because it changes direction of moving after reflection, instead of two times reflected wave that will be rejected on receiver input. Two different sensors were used to check effect of multi-passing at real conditions on railways.

At the first stage, the system was tested on unmoveable objects as a wall and a train. Also we get signal from a man is walking on railways.

The next test was done into actual moving of engine on railway. Signals from straight way, tunnel and some bends.

Results of these tests are presented pictures below. There are two pictures for each test: one is photo from digital camera and the next one is result of sensor measurements - plot of IF signal (white colour) and corresponding FFT function (green colour). As you can see on signal plot, there are signals from objects really presented on the picture. Based on this information we can measure distance to objects.

Due to achieved results we can do the following primary conclusions:

During field tests good operation of sensors were demonstrated. Sensor sensitivity and antenna pattern diagram are enough for the using in GAMAV project. On the other hand there are numbers of tasks which should be solved for final system. There is potentially possibility to measure both: speed of objects and distance at the same time. It is necessary to design new fast data processing algorithm for sensor to realise this feature. Two sensors or probably either more will be needed to achieve angle resolution specified in GAMAV. That will require significant R&D and engineering. To select signals at turns from objects around railway GPS intellectual system has to be used in parallel with sensor.