The Research On Technologies Of Route Query With User Activities

At present activity trajectory query is a popular research topic of location-based services.Activity trajectory is a spatial trajectory associated with activity keywords.And activity trajectory query algorithms include the traditional trajectory query and spatial keywords query.It allows users input several target locations as query locations and activity keywords as query condition to search activity trajectories which satisfy users' requirements.But these algorithms concentrate on textual matching and ignore the spatial limit.Although activity trajectories obtained by these algorithms can completely match to users' activity requirements,they can only get close to a part of query points and stay away from the other part of query locations due to activity keyword matching requirement and practical reasons.In fact,people need routes which near to query locations,but returned trajectories by these algorithms deviate from users' needs.Aiming at overcoming the above-mentioned shortcoming,this paper studies the match problem of activity trajectories and query locations,and proposes a method which uses trajectory segments to solve it.The method can divide entire activity trajectories into activity trajectory segments and combine trajectory segments which matched to query locations partly into a complete trajectory for users.Based on the above-mentioned idea,we propose two activity trajectory search algorithms,called as Ordered-match Activity Trajectory Segments Search(OATSS)and Activity Trajectory Segments Exploration(ATSE),to solve the problem about too large spatial matching distance between activity trajectories and query locations.The first algorithm,OATSS,searches activity trajectory segments which matched to query locations partly near to query locations and try to combine these segments into a complete trajectory.If the complete trajectory can match to the whole user query,it would be put into a candidate result set.At last,the algorithm retrieves top-k trajectories as results returned to users.The second algorithm,ATSE,begins with activity trajectory segments which near the first query location.It absorbs segments which matched to the next query locations and turns into a longer trajectory.In the searching process,if the trajectory after extension can match to user query completely,it would be put into a candidate result set.At last,the algorithm retrieves topk trajectories as results returned to users.Finally,we design several comparative experiments to test our proposed algorithms.We use both a real trajectory dataset and a synthetic trajectory dataset to compare two algorithms with a traditional activity trajectory algorithm under different parameter conditions.The experiment results demonstrate that two algorithms perform better than the traditional activity trajectory algorithm.And OATSS runs faster a little than ATSE,but results retrieved from ATSE are closer to query locations than OATSS.