One of the main difficulties I can think of would be the amount of emitted photons that would actually reach the scatterer without losing any energy or changing its course. These problems often occur due to elements in the environment of the experiment which should be large as we need an area for the train to reach the speed of sound.
Another practical difficulty would be reaching the speed of sound without hurting the setup containing the source.
A way to make this feasible is to determine a smaller area where the train reaches the speed of sound. This area should then be sufficiently covered to avoid environmental issues and have a decently wide range of scatterers in order to contain a select range of angles.