Research On Bump Features Of Road Surface Detection Method Based On Time Series Analysis Of Smartphone Sensor Streaming Data

Geographic information,as a fundamental surveying and mapping geographic data,provides powerful data support for the healthy and sustainable development of the national economy,and has extensive application value in many industry fields,including urban planning,traffic management,public safety,and disaster emergency response.Road data,as one of the key components of geographic information,mainly includes information on the basic attributes,traffic conditions,and passing capacity of roads.These data can help people better utilize the structure and function of the road network,and provide important references for urban traffic planning and management.Among them,the bump features of the road surface can express the use and damage of the road,and are of great significance for the maintenance of the road.Currently,there are mainly three methods for detecting bump features of the road surface: methods based on professional equipment,methods based on images,and methods based on sensors.The method based on professional equipment depends on professional equipment and operators,and has low detection efficiency.The method based on images is easily affected by weather and lighting factors,while the method based on sensors usually requires a specially designed or modified sensor carrier platform,and has limited popularity.This makes it difficult for current mainstream methods to quickly extract bump features of the road surface from massive data.Therefore,this paper proposes a non-smooth surface feature detection method based on time-series analysis of smartphone sensor stream data.This method uses the built-in sensors of smartphones to obtain time-series data and extract bump features of the road surface from it,providing a new idea for obtaining bump features of the road surface.Based on this,the paper conducts research in the following three aspects:(1)The basic principle of detecting bump features on road surfaces involves utilizing built-in sensors within smartphones to carry out the detection: The detection of bump features of the road surface using the built-in sensors of smartphones mainly involves the use of the following sensors:the accelerometer sensor used to detect the three-axis acceleration of the vehicle,the position sensor used to locate bump features of the road surface,and other sensors that assist in detecting bump features of the road surface.When a car with a smartphone onboard passes over bump features of the road surface,the three-axis acceleration of the car will undergo significant changes.By identifying the changes in sensor data,bump features of the road surface can be detected.(2)Modeling Bump Features of Road Surfaces: This article uses deep learning neural networks to detect bump features of the road surface in sensor time-series data.A neural network model for detecting bump features is designed,and the modeled data is input into the network for training.The output includes the sensor data and their geographic labels obtained by the neural network model.The geographic labels are then converted into bump features and the data is projected onto a map to obtain the final detection results.(3)Deep Learning-Based Detection Method: This paper adopts training deep learning neural networks to detect the bump features of the road surface in the sensor time series data.A neural network model is designed to detect the bump features,and the modeled data is input into the neural network for training.The trained neural network model outputs the sensor data and their geographical labels.The geographical labels are converted into bump features of the road surface,and the data is projected onto the map to obtain the final detection result.This paper conducts experiments on the data collected by smartphone sensors and analyzes long time-series sensor data using the proposed method to achieve the detection of bump features of the road surface based on smartphone sensor stream data.The experimental results show that the proposed method can quickly and accurately analyze the sensor stream data and extract the bump features of the road surface with a detection accuracy of over 85%,demonstrating high efficiency and accuracy.The method presented in this paper can detect bump features of the road surface,which is of great significance for obtaining road information and updating and maintaining basic geographic information data.