ebook ebooks e-book e-books downloaden bei MyEbooks.ch downloaden

Sensing Vehicle Conditions for Detecting Driving Behaviors

:	Jiadi Yu, Yingying Chen, Xiangyu Xu
:	Sensing Vehicle Conditions for Detecting Driving Behaviors
:	Springer-Verlag
:	9783319897707
:	1
:	CHF 48.30
:

:	Datenkommunikation, Netzwerke
:	English

:	81
:	Wasserzeichen/DRM
:	PC/MAC/eReader/Tablet
:	PDF

This SpringerBrief begins by introducing the concept of smartphone sensing and summarizing the main tasks of applying smartphone sensing in vehicles. Chapter 2 describes the vehicle dynamics sensing model that exploits the raw data of motion sensors (i.e., accelerometer and gyroscope) to give the dynamic of vehicles, including stopping, turning, changing lanes, driving on uneven road, etc. Chapter 3 detects the abnormal driving behaviors based on sensing vehicle dynamics. Specifically, this brief proposes a machine learning-based fine-grained abnormal driving behavior detection and identification system, D3, to perform real-time high-accurate abnormal driving behaviors monitoring using the built-in motion sensors in smartphones.

As more vehicles taking part in the transportation system in recent years, driving or taking vehicles have become an inseparable part of our daily life. However, increasing vehicles on the roads bring more traffic issues including crashes and congestions, which make it necessary to sense vehicle dynamics and detect driving behaviors for drivers. For example, sensing lane information of vehicles in real time can be assisted with the navigators to avoid unnecessary detours, and acquiring instant vehicle speed is desirable to many important vehicular applications. Moreover, if the driving behaviors of drivers, like inattentive and drunk driver, can be detected and warned in time, a large part of traffic accidents can be prevented. However for sensing vehicle dynamics and detecting driving behaviors, traditional approaches are grounded on the built-in infrastructure in vehicles such as infrared sensors and radars, or additional hardware like EEG devices and alcohol sensors, which involves high cost. The authors illustrate that smartphone sensing technology, which involves sensors embedded in smartphones (including the accelerometer, gyroscope, speaker, microphone, etc.), can be applied in sensing vehicle dynamics and driving behaviors.

Ch pter 4 exploits the feasibility to recognize abnormal driving events of drivers at early stage. Specifically, the authors develop an Early Recognition system, ER, which recognize inattentive driving events at an early stage and alert drivers timely leveraging built-in audio devices on smartphones. An overview of the state-of-the-art research is presented in chapter 5. Finally, the conclusions and future directions are provided in Chapter 6.

	Preface	6
	Contents	8
	1 Overview	10
	1.1 Brief Introduction of Smartphone Sensing	10
	1.1.1 Representative Sensors Embedded in Smartphones	10
	1.1.2 Development of Smartphone Sensing	11
	1.2 Smartphone Sensing in Vehicles	12
	1.3 Overview of the Book	13
	2 Sensing Vehicle Dynamics with Smartphones	15
	2.1 Introduction	15
	2.2 Pre-processing Sensor Readings	16
	2.2.1 Coordinate Alignment	16
	2.2.2 Data Filtering	18
	2.3 Sensing Basic Vehicle Dynamics	19
	2.3.1 Sensing Movement of Vehicles	19
	2.3.2 Sensing Driving on Uneven Road	20
	2.3.3 Sensing Turning of Vehicles	21
	2.3.4 Sensing Lane-Changes of Vehicles	22
	2.3.4.1 Identifying Single Lane-Change	22
	2.3.4.2 Identifying Sequential Lane-Change	23
	2.3.5 Estimating Instant Speed of Vehicles	25
	2.4 Evaluation	28
	2.4.1 Setup	28
	2.4.2 Metrics	28
	2.4.3 Performance of Sensing Vehicle Dynamics	29
	2.4.4 Performance of Sensing Lane-Change	29
	2.4.5 Performance of Sensing Instance Speed	30
	2.5 Conclusion	31
	3 Sensing Vehicle Dynamics for Abnormal Driving Detection	32
	3.1 Introduction	32
	3.2 Driving Behavior Characterization	35
	3.2.1 Collecting Data from Smartphone Sensors	35
	3.2.2 Analyzing Patterns of Abnormal Driving Behaviors	36
	3.3 System Design	37
	3.3.1 Overview	37
	3.3.2 Extracting and Selecting Effective Features	39
	3.3.2.1 Feature Extraction	39
	3.3.2.2 Feature Selection	39
	3.3.3 Training a Fine-Grained Classifier Model to Identify Abnormal Driving Behaviors	40
	3.3.4 Detecting and Identifying Abnormal Driving Behaviors	42
	3.4 Evaluations	44
	3.4.1 Setup	44
	3.4.2 Metrics	45
	3.4.3 Overall Performance	45
	3.4.3.1 Total Accuracy	45
	3.4.3.2 Detecting the Abnormal vs. the Normal	46
	3.4.3.3 Identifying Abnormal Driving Behaviors	46
	3.4.4 Impact of Training Set Size	47
	3.4.5 Impact of Traffic Conditions	48
	3.4.6 Impact of Road Type	48
	3.4.7 Impact of Smartphone Placement	49
	3.5 Conclusion	50
	4 Sensing Driver Behaviors for Early Recognition of Inattentive Driving	51
	4.1 Introduction	51
	4.2 Inattentive Driving Events Analysis	52
	4.2.1 Defining Inattentive Driving Events	53
	4.2.2 Analyzing Patterns of Inattentive Driving Events	54
	4.3 System Design	56
	4.3.1 System Overview	56
	4.3.2 Model Training at Offline Stage	57
	4.3.2.1 Establishing Training Dataset	57
	4.3.2.2 Extracting Effective Features	57
	4.3.2.3 Training a Multi-Classifier	58
	4.3.2.4 Setting Up Gradient Model Forest for Early Recognition	60
	4.3.3 Recognizing Inattentive Driving Events at Online Stage	62
	4.3.3.1 Segmenting Frames Through Sliding Window	62
	4.3.3.2 Detecting Inattentive Driving Events at Early Stage	63
	4.4 Evaluation	64
	4.4.1 Setup	64
	4.4.2 Metrics	64
	4.4.3 Overall Performance	65
	4.4.3.1 Total Accuracy	65
	4.4.3.2 Recognizing Inattentive Driving Events	66
	4.4.3.3 Realizing Early Recognition	66
	4.4.4 Impact of Training Set Size	67
	4.4.5 Impact of Road Types and Traffic Conditions	68
	4.4.6 Impact of Smartphone Placement	69
	4.5 Conclusion	69
	5 State-of-Art Researches	71
	5.1 Smartphone Sensing Researches	71
	5.2 Vehicle Dynamics Sensing Researches	72
	5.3 Driver Behaviors Detection Researches	73
	5.4 Common Issues	74
	6 Summary	75
	6.1 Conclusion of the Book	75
	6.2 Future Directions	76
	References	77