Added eventCollector to save events from speech and navigation

[vvasco]

This PR adds eventCollector, which relies on jsoncpp library to save data from speech and obstacle detection, in the following format:

{
    "Trial" : 
     [         
         {
             "Number": 0,
             "User-id": "skeleton-tag"        
             "Speech" : 
             {
                 "error-messages" : 
                 [ 
                     {
                          "google-speech": "e1",
                          "google-speech-event-time": "t1",
                          "google-speech-process": "e1process",
                          "google-speech-process-event-time": "t1process"
                     },
                     {
                           "google-speech": "e2",
                           "google-speech-event-time": "t2",
                           "google-speech-process": "e2process",
                           "google-speech-process-event-time": "t2process"
                      }
                 ]
                 },
             "Navigation" : 
             {
                 "error-messages" : 
                 [
                     {
                           "obstacle-detection-event-time": "t1",
                           "obstacle-detection": "obstacle found from front/rear laser at distance d"
                     },
                 ]
                 }
             }
       ]
 }

I left it draft as it should be merged after #311 (feat/etapas-speech is on top of fix/etapas-demo)

This closes #307 and closes #308.

Once this is merged, we need to rebuild the docker images and test the full through the compose files.

[pattacini]

Hi @vvasco

I've rebased the feat branch on top of the latest master and fixed the conflicts generated by my amendments to the skeletonRetriever.

Could you give a quick look at the present status of the PR, just to exclude possible mistakes I may have done? As soon as I have your ok, I'll be happy to merge.