Suttipong Thajchayapong

Detection of Driving Events using Sensory Data on Smartphone

In a fast-paced environment of today society, safety issue related to driving is considered a second priority in contrast to travelling from one place to another in the shortest possible time. This often leads to possible accidents. In order to reduce road traffic accidents, one domain which requires to be focused on is driving behaviour. This paper proposes three algorithms which detect driving events using motion sensors embedded on a smartphone since it is easily accessible and widely available in the market. More importantly, the proposed algorithms classify whether or not these events are aggressive based on raw data from various on board sensors on a smartphone. In addition, one of the outstanding features of the proposed algorithm is the ability to fine tune and adjust its sensitivity level to suit any given application domain appropriately. Initial experimental results reveal that the pattern matching algorithm outperforms the rule-based algorithm for driving events in both lateral and longitudinal movements where a high percentage of detection rate has been obtained for 11 out of 12 types of driving events. In addition, a trade-off between the detection rate and false alarm rate has been demonstrated under a range of algorithm settings in order to illustrate the proposed algorithm’s flexibility.

Driver behaviour profiling using smartphone sensory data in a V2I environment

Road traffic accidents prevention and mitigation are important issues that appear on top of the priority list of many countries around the world today. Many measures and approaches have been put in place in terms of policy level as well as technical level. Driver behaviour is one of many key factors that should be seriously considered to improve road safety. This paper proposes a method for driver behaviour profiling using sensory data on smartphones in a vehicle-to-infrastructure environment. Based on driving behaviours with the most risk to causing accidents, the profiling algorithm takes into account sudden driving events which occur during a journey to categorise drivers into different profiles according to their safety levels. The profiling algorithm offers the flexibility to adjust the parameters weightings in order to put an emphasis on specific driving events for different scenarios and applications. The impact on vehicle-to-infrastructure is that the stored driving profiles can be used to generate a norm for a given road section. Approaching vehicles deviating from the norm can be notified in real-time. Moreover, localised dangerous driving events can be clustered together to form a potential blackspot which can be deployed as an advanced warning for approaching vehicles as a location based service. As a result, the risk of road traffic accidents can be reduced. Real-world driving data was collected over two major routes in Thailand with four distinct profiles and five major factors to road accidents.

Traffic incident detection system using series of point detectors

This paper proposed a traffic incident detection system that can report the occurrence of traffic incidents, which occur between detectors. Based on the previously proposed dynamic time warping algorithm [2,3], this incident detection system monitors and assesses changes at upstream and downstream sites. Then, if the upstream-downstream changes are associated with traffic incidents, the system raised an alarm and report CCTV images on site, which are sent to traffic operators to further response. Performance evaluations are conducted using real-world traffic data where it is shown that the incident detection algorithm used in the proposed system achieves 94% detection rate and low false alarm rate. We also show that the proposed incident detection system outperforms a previously proposed incident detection system algorithm [4].

Combining Unsupervised Anomaly Detection and Neural Networks for Driver Identification

This paper proposes an algorithm for real-time driver identification using the combination of unsupervised anomaly detection and neural networks. The proposed algorithm uses nonphysiological signals as input, namely, driving behavior signals from inertial sensors (e.g., accelerometers) and geolocation signals from GPS sensors. First anomaly detection is performed to assess if the current driver is whom he/she claims to be. If an anomaly is detected, the algorithm proceeds to find relevant features in the input signals and use neural networks to identify drivers. To assess the proposed algorithm, real-world data are collected from ten drivers who drive different vehicles on several routes in real-world traffic conditions. Driver identification is performed on each of the seven-second-long driving behavior signals and geolocation signals in a streaming manner. It is shown that the proposed algorithm can achieve relatively high accuracy and identify drivers within 13 seconds. The proposed algorithm also outperforms the previously proposed driver identification algorithms. Furthermore, to demonstrate how the proposed algorithm can be deployed in real-world applications, results from real-world data associated with each operation of the proposed algorithm are shown step-by-step.

A Streaming Algorithm for Online Estimation of Temporal and Spatial Extent of Delays

Knowing traffic congestion and its impact on travel time in advance is vital for proactive travel planning as well as advanced traffic management. This paper proposes a streaming algorithm to estimate temporal and spatial extent of delays online which can be deployed with roadside sensors. First, the proposed algorithm uses streaming input from individual sensors to detect a deviation from normal traffic patterns, referred to as anomalies, which is used as an early indication of delay occurrence. Then, a group of consecutive sensors that detect anomalies are used to temporally and spatially estimate extent of delay associated with the detected anomalies. Performance evaluations are conducted using a real-world data set collected by roadside sensors in Bangkok, Thailand, and the NGSIM data set collected in California, USA. Using NGSIM data, it is shown qualitatively that the proposed algorithm can detect consecutive occurrences of shockwaves and estimate their associated delays. Then, using a data set from Thailand, it is shown quantitatively that the proposed algorithm can detect and estimate delays associated with both recurring congestion and incident-induced nonrecurring congestion. The proposed algorithm also outperforms the previously proposed streaming algorithm.

Budget and procurement analytics using open government data in Thailand

This paper proposes an open government data (OGD) analytics system for discovering patterns in budget requests and procurement data of government agencies. The proposed OGD analytics system is part of an ongoing project to pioneer possible use cases of OGD in Thailand. Experiments are conducted with budget and procurement data of government agencies for the fiscal years of 2013–2017. Part of our results have already been publicized on the website of Thailands Bureau of the Budget. Furthermore, four interesting patterns of government spending are found in the data and their possible implications are discussed. We anticipate that this paper will serve as an example on how government data in Thailand can be used to support e-government strategies to further deliver information to citizens.

Severity assessment of anomalies using driving behaviour signals

This paper reports an ongoing development of an algorithm for severity assessment of road anomalies using only driving behaviour signals. This algorithm is to be used in an active safety system where severity assessment operates in a distributed manner, i.e. at the leading vehicle that encounters an anomaly. Based on real-world 1,300 braking signals collected from 52,000 participants, it is shown that braking signals themselves can be clustered according to their duration and magnitude. These preliminary results demonstrate that it is feasible to cluster severity using driving behaviour signals.