MATLAB多传感器融合--核心步骤

MATLAB多传感器集成-核心步骤

MATLAB多传感器集成的核心步骤是stepImpl在函数中，该函数输入的跟踪目标和测量目标的信息输出为确认的真实目标信息和测试的跟踪目标信息。

[confirmedTracks, tentativeTracks, allTracks] = stepImpl(tracker, detections, varargin)

函数涉及的核心步骤如下所示

(1)通过成本矩阵计算测量的目标与跟踪目标相关，即Assignments 没有匹配的跟踪目标是什么？UnassignedTracks，哪些测量目标没有成功匹配？UnassignedDetections。

[Assignments, UnassignedTracks, UnassignedDetections] =...

tracker.associateDetectionsToTracksByUserCost(varargin{end});

(2) 将未匹配成功的测量目标UnassignedDetections初始化为新的跟踪目标

% Initiate tracks based on unassigned detections

initiateTracks(tracker, UnassignedDetections);

(3) 匹配成功的目标Assignments在跟踪目标中更新数据

% Update tracks based on assigned detections.

updateAssignedTracks(tracker, Assignments);

(4) 未成功匹配的跟踪目标UnassignedTracks根据时间决定是否删除

% Delete tracks that are not assigned

deleteOldTracks(tracker, UnassignedTracks, time);

(5)预测所有跟踪目标的状态

% Predict all the tracks to the end of the time step.

predictTracks(tracker, time);