Object recognition device and object recognition method

An object recognition device, which is configured to receive, from a sensor, a detection data group formed of a plurality of pieces of object data having the same associated time, to thereby generate pieces of track data on respective objects individually corresponding to the plurality of pieces of object data included in the received detection data group, the object recognition device comprising:

• a data receiver configured to receive the detection data group from the sensor, and associate the associated time with the received detection data group;
• a prediction processor configured to generate, for each of the objects, prediction data being a predicted value of each of the pieces of track data for the associated time;
• an association processor configured to associate, for the respective objects, the pieces of object data for the associated time and the pieces of prediction data for the associated time individually with each other;
• and
• an update processor configured to generate, for each of the objects, the track data for the associated time through use of the object data and the prediction data that correspond to each other for the associated time,
• wherein the association processor is configured to:
• classify, under a state in which not all the pieces of object data in the detection data group have been received and a part of the pieces of object data have been received by a current processing time, the part of the pieces of object data into pieces regarded as ready for association determination and pieces not regarded as ready for the association determination, to thereby temporarily associate the pieces of object data regarded as ready for the association determination and the pieces of prediction data individually with each other;
• and
• associate, under a state in which remaining pieces of object data in the detection data group have been received by the next processing time after the current processing time, all the pieces of object data in the detection data group and the pieces of prediction data individually with each other.

An object recognition method for receiving, from a sensor, a detection data group formed of a plurality of pieces of object data having the same associated time, to thereby generate pieces of track data on respective objects individually corresponding to the plurality of pieces of object data included in the received detection data group, the object recognition method comprising the steps of:

• receiving the detection data group from the sensor, and associating the associated time with the received detection data group;
• generating, for each of the objects, prediction data being a predicted value of each of the pieces of track data for the associated time;
• associating, for the respective objects, the pieces of object data for the associated time and the pieces of prediction data for the associated time individually with each other;
• and
• generating, for each of the objects, the track data for the associated time through use of the object data and the prediction data that correspond to each other for the associated time,
• wherein the step of associating includes:
• classifying, under a state in which not all the pieces of object data in the detection data group have been received and a part of the pieces of object data have been received by a current processing time, the part of the pieces of object data into pieces regarded as ready for association determination and pieces not regarded as ready for the association determination, to thereby associate the pieces of object data regarded as ready for the association determination and the pieces of prediction data individually with each other and set, as pieces of pending object data, the pieces of object data not regarded as ready for the association determination;
• and
• associating, under a state in which remaining pieces of object data in the detection data group have been received by the next processing time after the current processing time, the remaining pieces of object data and the pieces of pending object data individually with the pieces of prediction data.

An object recognition method for receiving, from a sensor, a detection data group formed of a plurality of pieces of object data having the same associated time, to thereby generate pieces of track data on respective objects individually corresponding to the plurality of pieces of object data included in the received detection data group, the object recognition method comprising the steps of:

• receiving the detection data group from the sensor, and associate the associated time with the received detection data group;
• generating, for each of the objects, prediction data being a predicted value of each of the pieces of track data for the associated time;
• associating, for the respective objects, the pieces of object data for the associated time and the pieces of prediction data for the associated time individually with each other;
• and
• generating, for each of the objects, the track data for the associated time through use of the object data and the prediction data that correspond to each other for the associated time,
• wherein the step of associating includes:
• classifying, under a state in which not all the pieces of object data in the detection data group have been received and a part of the pieces of object data have been received by a current processing time, the part of the pieces of object data into pieces regarded as ready for association determination and pieces not regarded as ready for the association determination, to thereby temporarily associate the pieces of object data regarded as ready for the association determination and the pieces of prediction data individually with each other;
• and
• associating, under a state in which remaining pieces of object data in the detection data group have been received by the next processing time after the current processing time, all the pieces of object data in the detection data group individually with the pieces of prediction data.