多假设跟踪算法（MHT）是一种在数据关联发生冲突时，形成多种假设以延迟做决定的逻辑。与PDA合并多种假设的做法不同，MHT算法把多个假设继续传递，让后续的观测数据解决这种不确定性。举个例子，PDA对所有假设以对应的概率进行加权平均，然后再对航迹进行更新。因此，如果有10个假设，PDA会将这10个假设有效的合并只留下一个假设。而另一方面，MHT却是保持这10个假设的子集并延迟决定，这样可以利用之后的观测数据解决当前扫描帧的不确定性问题。 -Multiple Hypothesis Tracking (MHT) is a kind of data association in the event of a conflict, the formation of a variety of assumptions in order to delay a decision logic. PDA combined with the practice of a variety of different assumptions, MHT algorithm is to pass on to a number of assumptions, so that follow-up observations to resolve this uncertainty. For example, PDA for all assumptions to the corresponding probability-weighted average, and then update the right track. Therefore, if there are 10 assumptions, PDA will be assumed that an effective merger of 10, leaving only a hypothesis. On the other hand, MHT was to keep this a subset of 10 hypothetical and delay the decision, so that after the observational data can be used to resolve the current scan fr a me uncertainties.
相关搜索： MHT
DATA
ASSOCIATION
AND
TRACKING
multiple
hypothesis
tracking
航迹
r214
tracking
and
data
association
PDA
MHT
algorithm
multiple
tracking
数据关联