Leon Stenneth - Profile on Academia.edu (original) (raw)

Papers by Leon Stenneth

Transportation Research Record, Jul 22, 2020

The state-of-the-art traffic sign recognition (TSR) algorithms are designed to recognize the text... more The state-of-the-art traffic sign recognition (TSR) algorithms are designed to recognize the textual information of a traffic sign at over 95% accuracy. Even though, they are still not ready for complex roadworks near ramps. In real-world applications, when the vehicles are running on the freeway, they may misdetect the traffic signs for the ramp, which will become inaccurate feedback to the autonomous driving applications and result in unexpected speed reduction. The misdetection problems have drawn minimal attention in recent TSR studies. In this paper, it is proposed that the existing TSR studies should transform from the point-based sign recognition to path-based sign learning. In the proposed pipeline, the confidence of the TSR observations from normal vehicles can be increased by clustering and location adjustment. A supervised learning model is employed to classify the clustered learned signs and complement their path information. Test drives are conducted in 12 European countries to calibrate the models and validate the path information of the learned sign. After model implementation, the path accuracy over 1,000 learned signs can be increased from 75.04% to 89.80%. This study proves the necessity of the path-based TSR studies near freeway ramps and the proposed pipeline demonstrates a good utility and broad applicability for sensor-based autonomous vehicle applications.

Transportation Research Record, 2020

Proceedings of the 1st international workshop on Mobile location-based service, 2011

Privacy in location-based systems is a major concern, since many mobile phones have a GPS sensor ... more Privacy in location-based systems is a major concern, since many mobile phones have a GPS sensor that can report location within 10 meters of accuracy. The contributions of this paper are in three folds. First, we examine privacy issues in snapshot queries, and present our work and results in this area. The proposed method can guarantee that all queries are protected, while previously proposed algorithms only achieve a low success rate in some situations. Next, we discuss continuous queries and illustrate that current snapshot solutions cannot be applied to continuous queries. Then, we present results for our robust models for continuous queries. Finally, we show evaluation results when we add another dimension to privacy in location based systems, referred to as transportation mode homogeneity.

Proceedings of the 19th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, 2011

The transportation mode such as walking, cycling or on a train denotes an important characteristi... more The transportation mode such as walking, cycling or on a train denotes an important characteristic of the mobile user's context. In this paper, we propose an approach to inferring a user's mode of transportation based on the GPS sensor on her mobile device and knowledge of the underlying transportation network. The transportation network information considered includes real time bus locations, spatial rail and spatial bus stop information. We identify and derive the relevant features related to transportation network information to improve classification effectiveness. This approach can achieve over 93.5% accuracy for inferring various transportation modes including: car, bus, aboveground train, walking, bike, and stationary. Our approach improves the accuracy of detection by 17% in comparison with the GPS only approach, and 9% in comparison with GPS with GIS models. The proposed approach is the first to distinguish between motorized transportation modes such as bus, car and aboveground train with such high accuracy. Additionally, if a user is travelling by bus, we provide further information about which particular bus the user is riding. Five different inference models including Bayesian Net, Decision Tree, Random Forest, Naïve Bayesian and Multilayer Perceptron, are tested in the experiments. The final classification system is deployed and available to the public.

Abstract. Many mobile phones have a GPS sensor that reports accurate location. Thus, if these loc... more Abstract. Many mobile phones have a GPS sensor that reports accurate location. Thus, if these location data are not protected adequately, they may cause privacy breeches. Several reports are available where people have been stalked through GPS. The contributions of this paper are in two folds. First, we examine privacy issues in snapshot queries and propose a method that guarantees that all queries are protected. Previously proposed algorithms achieve a low success rate in some situations. Next, we study continuous ...

Machine Learning and Knowledge Discovery in Databases

Estimating traffic conditions in arterial networks with GPS probe data is a practically important... more Estimating traffic conditions in arterial networks with GPS probe data is a practically important while substantially challenging problem. With the increasing availability of GPS equipments installed in various vehicles, GPS probe data is currently becoming a significant data source for traffic monitoring. However, limited by the lack of reliability and low sampling frequency of GPS probes, probe data are usually not sufficient for fully estimating traffic conditions of a large arterial network. For the first time this paper studies how to explore social media as an auxiliary data source and incorporate it with GPS probe data to enhance traffic congestion estimation. Motivated by the increasing amount of traffic information available in Twitter, we first extensively collect tweets that report various traffic events such as congestion, accident, and road construction. Next we propose an extended Coupled Hidden Markov Model which can effectively integrate GPS probe readings and traffic related tweets to more accurately estimate traffic conditions of an arterial network. To address the computational challenge, a sequential importance sampling based EM algorithm is also introduced. We evaluate the proposed model on the arterial network of downtown Chicago. The experimental results demonstrate the superior performance of the model by comparison with previous methods.

Detecting Human Activities Using Smartphones and Maps

Parking Payment Detection

Method and apparatus for transit mapping

Turn restriction determination

Automated transportation transfer detection using GPS enabled smartphones

Abstract Understanding the mobility of a traveller from mobile sensor data is an important area o... more Abstract Understanding the mobility of a traveller from mobile sensor data is an important area of work in context aware and ubiquitous computing. Given a multimodal GPS trace, we will identify where in the GPS trace the traveller changed transportation modes. For example, where in the GPS trace the traveller alight a bus and boards a train, or where did the client stop running and start walking.

Users of mass transit systems such as those of buses and trains normally rely on accurate route m... more Users of mass transit systems such as those of buses and trains normally rely on accurate route maps, stop locations, and service schedules when traveling. If the route map, service schedule, or stop location has errors it can reduce the transit agency’s ridership. In this paper, the problem of deriving transit systems by mining raw GPS data is studied. Specifically, we propose and evaluate novel classification features with spatial and temporal clustering techniques that derive bus stop locations, route geometries, and service schedules from GPS data. Subsequently, manual and expensive field visits to record and annotate the initial or updated route geometries, transit stop locations, or service schedules is no longer required by transit agencies. This facilitates a massive reduction in cost for transit agencies. The effectiveness of the proposed algorithms is validated on the third largest public transit system in the United States.

Real-time parking availability information is important in urban areas, and if available could re... more Real-time parking availability information is important in urban areas, and if available could reduce congestion, pollution, and gas consumption. In this paper, we present a software solution called PhonePark for detecting the availability of on-street parking spaces. The solution uses the GPS and/or accelerometer sensors in a traveler’s mobile phone to automatically detect when and where the traveler parked her car, and when she released a parking slot. PhonePark can also utilize the mobile phone’s Bluetooth sensor or piggyback on street parking payment transactions for parking activity detection. Thus, the solution considers only mobile phones and does not rely on any external sensors such as cameras, wireless sensors embedded in the pavements, or ultrasonic sensors on vehicles. Further contributions include an algorithm to compute the historical parking availability profile for an arbitrary street block and algorithms to estimate the parking availability in real-time for a given street block. The algorithms are evaluated using real-time and real world street parking data.

Activity recognition from smartphone sensor traces and geospatial knowledge

Determining a traveler's context from wearable computers is essential in pervasive computing and ... more Determining a traveler's context from wearable computers is essential in pervasive computing and activity recognition. In this paper, we study activity recognition using mobile phones. More specifically, we propose a supervised learning scheme that uses a traveler's GPS and accelerometer sensors on their mobile phones, and geographic knowledge of the underlying transportation network, to detect the transportation mode that she is using in real time.

Real-time street parking availability information is important in urban areas, and if available c... more Real-time street parking availability information is important in urban areas, and if available could reduce congestion, pollution, and gas consumption. In this paper, an advanced street parking system called PhonePark is presented. Using the GPS, accelerometer, and Bluetooth sensors on a traveler's mobile phone, in conjunction with geospatial data, we can automatically detect when and where the traveler parked her car, and when she released a parking slot. 1

The transportation mode such as walking, cycling or on a train denotes an important characteristi... more The transportation mode such as walking, cycling or on a train denotes an important characteristic of the mobile user’s context. In this paper, we propose an approach to infer a user’s mode of transportation based on the GPS sensors on their mobile devices, and knowledge of the underlying transportation network. The transportation network information considered includes real time bus locations, spatial rail and spatial bus stop information. We identify and derive the relevant features related to transportation network information to improve classification effectiveness. This approach can achieve over 93.5% accuracy for inferring various transportation modes including: car, bus, aboveground train, walking, bike, and stationary. Compared with existing methods, our approach improves the accuracy of detection by 17% for GPS only approach, and 9% for GPS with GIS models. This proposed technique is the first to distinguish between motorized transportation modes such as bus, car and aboveground train with such high accuracy. Additionally, if a user is travelling by bus, we provide further information about which particular bus the user is travelling on. Five different inference models including, Bayesian Net, Decision Tree, Random Forest, Naïve Bayesian and Multilayer Perceptron, are studied in the experiments. The final classification system is deployed and available to the public.

A major concern for deployment of location-based mobile systems is the ill-usage of mobile client... more A major concern for deployment of location-based mobile systems is the ill-usage of mobile client’s location data, which may imply sensitive and private personal information. Also, even if the location is exposed willingly by the mobile client the query should not be linked to the mobile client. Still, many location based systems (store finders, transit itinerary systems, and social networks) are created with a different focus and have little concern for end users privacy. We focused on location based mobile systems where the location of the mobile user is available; however, an adversary should not be able to link a query to a specific mobile user. Two key contributions of this work are the introduction and experimental evaluation of a novel concept called transportation mode homogeneity anonymization that adds another dimension to privacy in mobile location based systems. Also, a novel dynamic layered approach on achieving k-anonymity by separating the local privacy requirement on each snapshot and global privacy requirement across snapshots with different privacy goals and exploits the local privacy anonymization group as candidates to obtain global anonymization group candidates.

Transportation Research Record, Jul 22, 2020

Transportation Research Record, 2020

Proceedings of the 1st international workshop on Mobile location-based service, 2011

Proceedings of the 19th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, 2011

Machine Learning and Knowledge Discovery in Databases

Detecting Human Activities Using Smartphones and Maps

Parking Payment Detection

Method and apparatus for transit mapping

Turn restriction determination

Automated transportation transfer detection using GPS enabled smartphones

Activity recognition from smartphone sensor traces and geospatial knowledge

The transportation mode such as walking, cycling or on a train denotes an important characteristi... more The transportation mode such as walking, cycling or on a train denotes an important characteristic of the mobile user’s context. In this paper, we propose an approach to infer a user’s mode of transportation based on the GPS sensors on their mobile devices, and knowledge of the underlying transportation network. The transportation network information considered includes real time bus locations, spatial rail and spatial bus stop information. We identify and derive the relevant features related to transportation network information to improve classification effectiveness. This approach can achieve over 93.5% accuracy for inferring various transportation modes including: car, bus, aboveground train, walking, bike, and stationary. Compared with existing methods, our approach improves the accuracy of detection by 17% for GPS only approach, and 9% for GPS with GIS models. This proposed technique is the first to distinguish between motorized transportation modes such as bus, car and aboveground train with such high accuracy. Additionally, if a user is travelling by bus, we provide further information about which particular bus the user is travelling on. Five different inference models including, Bayesian Net, Decision Tree, Random Forest, Naïve Bayesian and Multilayer Perceptron, are studied in the experiments. The final classification system is deployed and available to the public.