Modeling And Assisted Positioning Of Spatial Relationship Uncertainty In Net Car's Passenger Locality Description

In recent years,online car-hailing market has risen rapidly.Online car-hailing takes good advantage of the mobile Internet to save resources and time for both drivers and passengers,which has become a very popular way for the public to hail a car.However,while using the online car-hailing there are often problems like the driver detours and the passenger waits for a long time,which all due to the inaccurate starting point.The cause for this problem is that the existing mobile phone positioning technology has strong dependence on the objective environment,and the positioning accuracy will be reduced in an environment with severe occlusion and poor visibility.The development of cognitive technology and natural language understanding makes the communicate between humans and machines more mature.The natural language interactive via voice or text is more in line with human cognitive habits,which will be the future direction of location based services(LBS).After receiving a order,driver of the online-hailing car often ask the passengers for the location by SMS or telephone.Locality description exists in the daily communication between the two parties of driver and passenger.Locality description generally includes three parts: reference objects,spatial relationships,and target objects,wherein the spatial position of the reference objects and the spatial relationships between the target objects and the reference objects can be used for positioning.In recent years,the positioning methods based on locality description has made great progress,mainly focusing on application and research under large-scale scales.In small scales,there have also been explorations such as positioning in indoor shopping market and pedestrian location navigation,but no research on positioning method based on locality description for online car-hailing application.Aiming at the problem that the passengers' mobile phone positioning are often inaccurate while using online car-hailing,this paper proposes an assisted positioning method based on passenger's locality description.The main idea is using the cognitive statistics law to model the uncertainty of the spatial relationship in the passenger's locality description,calculate the uncertainty area and uncertainty of the passenger's position and obtain the reasonable range of passenger's position accordingly.Using the range to constrain the mobile phone positioning result,when the mobile phone positioning result exceeds the range,the mobile positioning result will be adjusted,so that the final positioning result conforms to the passenger's position description,so as to achieve the purpose of reducing the positioning error of the mobile phone under the condition that the GPS signal is severely blocked.The main research work and innovations of this paper are as follows:(1)The situational characteristics of the online car-hailing and the characteristics of the passengers' locality description while wating fort the car are annalyzed;And the main locality description patterns while wating for a online-hailing car are summarized.(2)Taking the subway station as an example,the location description experiment and the quantitative distance cognition experiment were carried out,and the quantitative distance model and the direction relationship model with the subway station as the reference object were established,which laid a foundation for the assisted positioning method of passenger's locality description,which can be used as a reference for uncertainty modeling of quantitative distance and direction relationships for other types of features in city.(3)An assisted positioning method using passenger's locality description based on the directional relationship and the quantitative distance relationship is proposed.The method constrains the passenger's position according to the directional relationship and the quantitative distance relationship in the passenger's locality description,and adjusts the mobile phone positioning result according to the constraint range.Experiments show that when the location mode of mobile phone is base station,the location description assistant location method can significantly improve the location accuracy,and reduce the average error from 144 m to 18.5m.The experimental results show that the position description assistant positioning method proposed in this paper can effectively improve the positioning accuracy of mobile phones when the environment is seriously occluded and the GPS signal is not good.