Support for Realsense D435 and D435i depth cameras

- change output type from depth_array to gray16_array, which allows us to use this output format for any 16bit grayscale image, not just depth array.  Based on pull request feedback from Rahul.

Voice control update

2. fix line length, grammatic and formatting issues
3. Error handling if request failed
4. Fix README

Voice control by Alexa

Fix issues with xboxswapped controller

- added a comment to clarify what dependencies are necessary and how to install them.

- Changed to allows use default time for wait_for_frames()
- Those above changes were verified in realsense_viewer.py in the realsenseutils project to make the 3 streams stable and reliable on the jetson nano

- added traceback in Vehicle.py to print any exception that terminates the vehicle loop, other than a keyboard interupt.

- added traceback in Vehicle.py to print any exception that terminates the vehicle loop, other than a keyboard interupt.

- added exception handler in Vehicle.py to print any exception that terminates the vehicle loop, other than a keyboard interupt.

- modified complete.py/manage.py to save depth and 'depth_array' type, so it can be handled specially in datastore.py
- modified datastore.py to handle 'depth_array' as a 16bit grayscale PNG image.

- give thread more time to quit.
- give camera more time to warm up.

- removed small 'i' from REALSENSE_D435i_IMU because it would not be copied with the lowercase letter in it, so I was getting an error about the config not existing.

Add option to do nothing on older axes, otherwise both joystics were
reacting in the same time to throttle and steering, which made it
totally unusable.

Path follow

RealSense-part added to the self test in realsense435i.py.  Added default config in cfg_complete.py.  Integrated into complete.py

Update faq.md

Update build_hardware.md

Arduino drivetrain support