Real-time decision making by driverless city vehicles : a discrete event driven approach

This thesis addresses the topic of real-time decision making by driverless (autonomous) city vehicles, i.e. their ability to make appropriate driving decisions in non-simplified urban traffic conditions. After addressing the state of research, and explaining the research question, the thesis presents solutions for the subcomponents which are relevant for decision making with respect to information input (World Model), information output (Driving Maneuvers), and the real-time decision making process. TheWorld Model is a software component developed to fulfill the purpose of collecting information from perception and communication subsystems, maintaining an up-to-date view of the vehicle’s environment, and providing the required input information to the Real-Time Decision Making subsystem in a well-defined, and structured way. The real-time decision making process consists of two consecutive stages. While the first decision making stage uses a Petri net to model the safetycritical selection of feasible driving maneuvers, the second stage uses Multiple Criteria Decision Making (MCDM) methods to select the most appropriate driving maneuver, focusing on fulfilling objectives related to efficiency and comfort. The complex task of autonomous driving is subdivided into subtasks, called driving maneuvers, which represent the output (i.e. decision alternatives) of the real-time decision making process. Driving maneuvers are considered as implementations of closed-loop control algorithms, each capable of maneuvering the autonomous vehicle in a specific traffic situation. Experimental tests in both a 3D simulation and real-world experiments attest that the developed approach is suitable to deal with the complexity of real-world urban traffic situations.

Tweeting the TV event, creating ‘Public Sphericules’ : ad hoc engagement with SBS’s Go Back to Where You Came From – season two

This article uses the example of the mediatisation of Season 2 of the Australian documentary-cum-reality TV series Go Back to Where You Came From, and the associated #GoBackSBS Twitter feed, to investigate how public opinions are shaped, reshaped and expressed in new hybrid media ecologies. We explore how social media tools like Twitter can support the efforts of a TV production; provide spaces through which the public can engage ad hoc with a public event, be informed, shape their opinions and share them with others; and thus open up new possibilities for public discourse to occur. We suggest that new online public sphericules are emerging that provide spaces within which publics can engage with the cultural social and political realities with which they are confronted. In this way, we highlight the importance of mundane communication to the shaping and constant reshaping of public opinion.

Using the Event Analysis of Systemic Teamwork (EAST) to explore conflict between different road user groups when making right hand turns at urban intersections

Collisions between different types of road users at intersections form a substantial component of the road toll. This paper presents an analysis of driver, cyclist, motorcyclist and pedestrian behaviour at intersections that involved the application of an integrated suite of ergonomics methods, the Event Analysis of Systemic Teamwork (EAST) framework, to on-road study data. EAST was used to analyse behaviour at three intersections using data derived from an on-road study of driver, cyclist, motorcyclist and pedestrian behaviour. The analysis shows the differences in behaviour and cognition across the different road user groups and pinpoints instances where this may be creating conflicts between different road users. The role of intersection design in creating these differences in behaviour and resulting conflicts is discussed. It is concluded that currently intersections are not designed in a way that supports behaviour across the four forms of road user studied. Interventions designed to improve intersection safety are discussed.

Unusual event detection in crowded scenes

Novel computer vision techniques have been developed to automatically detect unusual events in crowded scenes from video feeds of surveillance cameras. The research is useful in the design of the next generation intelligent video surveillance systems. Two major contributions are the construction of a novel machine learning model for multiple instance learning through compressive sensing, and the design of novel feature descriptors in the compressed video domain.

Near-miss event detection at railway level crossings

Recent modelling of socio-economic costs by the Australian railway industry in 2010 has estimated the cost of level crossing accidents to exceed AU$116 million annually. To better understand the causal factors of these accidents, a video analytics application is being developed to automatically detect near-miss incidents using forward facing videos from trains. As near-miss events occur more frequently than collisions, by detecting these occurrences there will be more safety data available for analysis. The application that is being developed will improve the objectivity of near-miss reporting by providing quantitative data about the position of vehicles at level crossings through the automatic analysis of video footage. In this paper we present a novel method for detecting near-miss occurrences at railway level crossings from video data of trains. Our system detects and localizes vehicles at railway level crossings. It also detects the position of railways to calculate the distance of the detected vehicles to the railway centerline. The system logs the information about the position of the vehicles and railway centerline into a database for further analysis by the safety data recording and analysis system, to determine whether or not the event is a near-miss. We present preliminary results of our system on a dataset of videos taken from a train that passed through 14 railway level crossings. We demonstrate the robustness of our system by showing the results of our system on day and night videos.

Stop-signal response inhibition in schizophrenia : Behavioural, event-related potential and functional neuroimaging data

Inhibitory control deficits are well documented in schizophrenia, supported by impairment in an established measure of response inhibition, the stop-signal reaction time (SSRT). We investigated the neural basis of this impairment by comparing schizophrenia patients and controls matched for age, sex and education on behavioural, functional magnetic resonance imaging (fMRI) and event-related potential (ERP) indices of stop-signal task performance. Compared to controls, patients exhibited slower SSRT and reduced right inferior frontal gyrus (rIFG) activation, but rIFG activation correlated with SSRT in both groups. Go stimulus and stop-signal ERP components (N1/P3) were smaller in patients, but the peak latencies of stop-signal N1 and P3 were also delayed in patients, indicating impairment early in stop-signal processing. Additionally, response-locked lateralised readiness potentials indicated response preparation was prolonged in patients. An inability to engage rIFG may predicate slowed inhibition in patients, however multiple spatiotemporal irregularities in the networks underpinning stop-signal task performance may contribute to this deficit.

Designing for grassroots food production : an event-based urban agriculture community

As urbanisation of the global population has increased above 50%, growing food in urban spaces increases in importance, as it can contribute to food security, reduce food miles, and improve people’s physical and mental health. Approaching the task of growing food in urban environments is a mixture of residential growers and groups. Permablitz Brisbane is an event-centric grassroots community that organises daylong ‘working bee’ events, drawing on permaculture design principles in the planning and design process. Permablitz Brisbane provides a useful contrast from other location-centric forms of urban agriculture communities (such as city farms or community gardens), as their aim is to help encourage urban residents to grow their own food. We present findings and design implications from a qualitative study with members of this group, using ethnographic methods to engage with and understand how this group operates. Our findings describe four themes that include opportunities, difficulties, and considerations for the creation of interventions by Human-Computer Interaction (HCI) designers.

An efficient and robust system for multi-person event detection in real world indoor surveillance scenes

Due to the popularity of security cameras in public places, it is of interest to design an intelligent system that can efficiently detect events automatically. This paper proposes a novel algorithm for multi-person event detection. To ensure greater than real-time performance, features are extracted directly from compressed MPEG video. A novel histogram-based feature descriptor that captures the angles between extracted particle trajectories is proposed, which allows us to capture motion patterns of multi-person events in the video. To alleviate the need for fine-grained annotation, we propose the use of Labelled Latent Dirichlet Allocation, a “weakly supervised” method that allows the use of coarse temporal annotations which are much simpler to obtain. This novel system is able to run at approximately ten times real-time, while preserving state-of-theart detection performance for multi-person events on a 100-hour real-world surveillance dataset (TRECVid SED).

The N170 observed ‘in the wild’: robust event-related potentials to faces in cluttered dynamic visual scenes

As a social species in a constantly changing environment, humans rely heavily on the informational richness and communicative capacity of the face. Thus, understanding how the brain processes information about faces in real-time is of paramount importance. The N170 is a high temporal resolution electrophysiological index of the brain's early response to visual stimuli that is reliably elicited in carefully controlled laboratory-based studies. Although the N170 has often been reported to be of greatest amplitude to faces, there has been debate regarding whether this effect might be an artifact of certain aspects of the controlled experimental stimulation schedules and materials. To investigate whether the N170 can be identified in more realistic conditions with highly variable and cluttered visual images and accompanying auditory stimuli we recorded EEG 'in the wild', while participants watched pop videos. Scene-cuts to faces generated a clear N170 response, and this was larger than the N170 to transitions where the videos cut to non-face stimuli. Within participants, wild-type face N170 amplitudes were moderately correlated to those observed in a typical laboratory experiment. Thus, we demonstrate that the face N170 is a robust and ecologically valid phenomenon and not an artifact arising as an unintended consequence of some property of the more typical laboratory paradigm.

An MRF based abnormal event detection approach using motion and appearance features

Abnormal event detection has attracted a lot of attention in the computer vision research community during recent years due to the increased focus on automated surveillance systems to improve security in public places. Due to the scarcity of training data and the definition of an abnormality being dependent on context, abnormal event detection is generally formulated as a data-driven approach where activities are modeled in an unsupervised fashion during the training phase. In this work, we use a Gaussian mixture model (GMM) to cluster the activities during the training phase, and propose a Gaussian mixture model based Markov random field (GMM-MRF) to estimate the likelihood scores of new videos in the testing phase. Further-more, we propose two new features: optical acceleration, and the histogram of optical flow gradients; to detect the presence of any abnormal objects and speed violations in the scene. We show that our proposed method outperforms other state of the art abnormal event detection algorithms on publicly available UCSD dataset.

Ranking based clustering for social event detection

The problem of clustering a large document collection is not only challenged by the number of documents and the number of dimensions, but it is also affected by the number and sizes of the clusters. Traditional clustering methods fail to scale when they need to generate a large number of clusters. Furthermore, when the clusters size in the solution is heterogeneous, i.e. some of the clusters are large in size, the similarity measures tend to degrade. A ranking based clustering method is proposed to deal with these issues in the context of the Social Event Detection task. Ranking scores are used to select a small number of most relevant clusters in order to compare and place a document. Additionally,instead of conventional cluster centroids, cluster patches are proposed to represent clusters, that are hubs-like set of documents. Text, temporal, spatial and visual content information collected from the social event images is utilized in calculating similarity. Results show that these strategies allow us to have a balance between performance and accuracy of the clustering solution gained by the clustering method.