Implement voice recognition
Implement basic simple commands such as: zoom, "zoom out", "show destination", "show current location", "turn on XXX" and "turn of XXX" where XX is a feature or setting.
Implement more complex and conceptual commands like: "How far is it to next gas-station", answer should be read out using the speech API and include
- Distance to gas station (time and km)
- Information on if it is along the current route or if it will be a de-tour.
Follow-up commands should be possible...
- add waypoint and recalculate route
- next on route
- go back to the previous one
- next one (the second closest not necessarily along the route)
- with rest-room (restaurant, cafe)
- with FUEL (Fuel could be diesel, gas, electricity/charging, gas, alcohol, ....)
Didn't say it would be easy, but it should be possible...you need to use Fuzzy Logic and hard code as little as possible and instead use POI, map-data and other sources to try figure out what the speaker wants.
Of course there should be some grammar that needs to be followed and some syntax rules.