Ransalu Senanayake

Ransalu Senanayake
Stanford University | SU · Department of Computer Science

Doctor of Philosophy

About

57
Publications
9,986
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
388
Citations
Introduction
With the aim of developing trustworthy AI systems, my research focuses on developing explainable machine learning algorithms for modeling spatiotemporal uncertainty.
Additional affiliations
July 2015 - June 2018
The University of Sydney
Position
  • Research Assistant
Description
  • Machine Learning and Data Mining
February 2015 - February 2019
The University of Sydney
Position
  • PhD Student
Description
  • Thesis: Nonlinear Methods for Spatiotemporal Mapping in Unstructured Environments
September 2013 - February 2015
Sri Lanka Institute of Information Technology
Position
  • Lecturer in Electrical and Computer Engineering
Education
February 2015 - January 2019
The University of Sydney
Field of study
  • Machine Learning and Robotics
February 2012 - August 2013
The Hong Kong University of Science and Technology
Field of study
  • Human-Computer Interaction
February 2009 - November 2011
Sheffield Hallam University
Field of study
  • Electronic Engineering

Publications

Publications (57)
Preprint
Full-text available
Safe and reliable autonomy solutions are a critical component of next-generation intelligent transportation systems. Autonomous vehicles in such systems must reason about complex and dynamic driving scenes in real time and anticipate the behavior of nearby drivers. Human driving behavior is highly nuanced and specific to individual traffic particip...
Preprint
Deep reinforcement learning (RL) is a promising approach to solving complex robotics problems. However, the process of learning through trial-and-error interactions is often highly time-consuming, despite recent advancements in RL algorithms. Additionally, the success of RL is critically dependent on how well the reward-shaping function suits the t...
Preprint
Characterizing aleatoric and epistemic uncertainty on the predicted rewards can help in building reliable reinforcement learning (RL) systems. Aleatoric uncertainty results from the irreducible environment stochasticity leading to inherently risky states and actions. Epistemic uncertainty results from the limited information accumulated during lear...
Preprint
Full-text available
We propose a unified framework for coordinating Unmanned Aerial Vehicle (UAV) swarms operating under time-varying communication networks. Our framework builds on the concept of graphical games, which we argue provides a compelling paradigm to subsume the interaction structures found in networked UAV swarms thanks to the shared local neighborhood pr...
Preprint
Full-text available
We generalize the derivation of model predictive path integral control (MPPI) to allow for a single joint distribution across controls in the control sequence. This reformation allows for the implementation of adaptive importance sampling (AIS) algorithms into the original importance sampling step while still maintaining the benefits of MPPI such a...
Preprint
Autonomous vehicles (AVs) rely on environment perception and behavior prediction to reason about agents in their surroundings. These perception systems must be robust to adverse weather such as rain, fog, and snow. However, validation of these systems is challenging due to their complexity and dependence on observation histories. This paper present...
Preprint
We present a method for autonomous exploration of large-scale unknown environments under mission time constraints. We start by proposing the Frontloaded Information Gain Orienteering Problem (FIG-OP) -- a generalization of the traditional orienteering problem where the assumption of a reliable environmental model no longer holds. The FIG-OP address...
Preprint
Full-text available
We present a novel approach to maximize the communication-aware coverage for robots operating over large-scale geographical regions of interest (ROIs). Our approach complements the underlying network topology in neighborhood selection and control, rendering it highly robust in dynamic environments. We formulate the coverage as a multi-stage, cooper...
Preprint
Many applications of generative models rely on the marginalization of their high-dimensional output probability distributions. Normalization functions that yield sparse probability distributions can make exact marginalization more computationally tractable. However, sparse normalization functions usually require alternative loss functions for train...
Preprint
Full-text available
Autonomous vehicles need to model the behavior of surrounding human driven vehicles to be safe and efficient traffic participants. Existing approaches to modeling human driving behavior have relied on both data-driven and rule-based methods. While data-driven models are more expressive, rule-based models are interpretable, which is an important req...
Preprint
Future urban transportation concepts include a mixture of ground and air vehicles with varying degrees of autonomy in a congested environment. In such dynamic environments, occupancy maps alone are not sufficient for safe path planning. Safe and efficient transportation requires reasoning about the 3D flow of traffic and properly modeling uncertain...
Preprint
Full-text available
Deep learning models have become a popular choice for medical image analysis. However, the poor generalization performance of deep learning models limits them from being deployed in the real world as robustness is critical for medical applications. For instance, the state-of-the-art Convolutional Neural Networks (CNNs) fail to detect adversarial sa...
Preprint
Predicting the future occupancy state of an environment is important to enable informed decisions for autonomous vehicles. Common challenges in occupancy prediction include vanishing dynamic objects and blurred predictions, especially for long prediction horizons. In this work, we propose a double-prong neural network architecture to predict the sp...
Preprint
Neural networks (NNs) are widely used for object recognition tasks in autonomous driving. However, NNs can fail on input data not well represented by the training dataset, known as out-of-distribution (OOD) data. A mechanism to detect OOD samples is important in safety-critical applications, such as automotive perception, in order to trigger a safe...
Preprint
Discrete latent spaces in variational autoencoders have been shown to effectively capture the data distribution for many real-world problems such as natural language understanding, human intent prediction, and visual scene representation. However, discrete latent spaces need to be sufficiently large to capture the complexities of real-world data, r...
Conference Paper
Full-text available
Creating accurate spatial representations that take into account uncertainty is critical for autonomous robots to safely navigate in unstructured environments. Although recent LIDAR based mapping techniques can produce robust occupancy maps, learning the parameters of such models demand considerable computational time, discouraging them from being...
Preprint
We can use driving data collected over a long period of time to extract rich information about how vehicles behave in different areas of the roads. In this paper, we introduce the concept of directional primitives, which is a representation of prior information of road networks. Specifically, we represent the uncertainty of directions using a mixtu...
Preprint
Full-text available
Creating accurate spatial representations that take into account uncertainty is critical for autonomous robots to safely navigate in unstructured environments. Although recent LIDAR based mapping techniques can produce robust occupancy maps, learning the parameters of such models demand considerable computational time, discouraging them from being...
Preprint
Full-text available
Imitation learning is an approach for generating intelligent behavior when the cost function is unknown or difficult to specify. Building upon work in inverse reinforcement learning (IRL), Generative Adversarial Imitation Learning (GAIL) aims to provide effective imitation even for problems with large or continuous state and action spaces. Driver m...
Preprint
Driver models are invaluable for planning in autonomous vehicles as well as validating their safety in simulation. Highly parameterized black-box driver models are very expressive, and can capture nuanced behavior. However, they usually lack interpretability and sometimes exhibit unrealistic-even dangerous-behavior. Rule-based models are interpreta...
Preprint
This paper addresses the problem of learning instantaneous occupancy levels of dynamic environments and predicting future occupancy levels. Due to the complexity of most real-world environments, such as urban streets or crowded areas, the efficient and robust incorporation of temporal dependencies into otherwise static occupancy models remains a ch...
Article
Full-text available
Autonomous vehicles operating in urban environments need to deal with an abundance of other dynamic objects, such as pedestrians and vehicles. This requires the development of predictive models that capture the complexity and long-term patterns of motion in the environment. We approach this problem by modelling movement directions of a typical obje...
Conference Paper
Full-text available
We present a spatio-temporal modelling method for robots operating in human-populated environments for extended time periods. The presented method integrates observations of pedestrians at different locations and times into an efficient representation of spatial and temporal structure of pedestrian flows. Long-term variations of the observed flows...
Conference Paper
Full-text available
Mapping the occupancy of an environment is central for robot autonomy. Traditional occupancy grid maps discretise the environment into independent cells, neglecting important spatial correlations, and are unable to capture the continuous nature of the real world. With these drawbacks of grid maps in mind, Hilbert Maps (HM) and more recently Bayesia...
Conference Paper
Full-text available
Kernel methods have been successfully used in various domains to model nonlinear patterns. However, the structure of the kernels is typically handcrafted for each dataset based on the experience of the data analyst. In this paper, we present a novel technique to learn kernels that best fit the data. We exploit the measure-theoretic view of a shift-...
Conference Paper
Full-text available
Robots often have to deal with the challenges of operating in dynamic and sometimes unpredictable environments. Although an occupancy map of the environment is sufficient for navigation of a mobile robot or manipulation tasks with a robotic arm in static environments, robots operating in dynamic environments demand richer information to improve rob...
Conference Paper
Full-text available
In order to deploy robots in previously unseen and unstructured environments, the robots should have the capacity to learn on their own and adapt to the changes in the environments. For instance, in mobile robotics, a robot should be able to learn a map of the environment from data itself without the intervention of a human to tune the parameters o...
Preprint
Full-text available
Robots often have to deal with the challenges of operating in dynamic and sometimes unpredictable environments. Although an occupancy map of the environment is sufficient for navigation of a mobile robot or manipulation tasks with a robotic arm in static environments, robots operating in dynamic environments demand richer information to improve rob...
Article
Mapping the occupancy level of an environment is important for a robot to navigate in unknown and unstructured environments. To this end, continuous occupancy mapping techniques which express the probability of a location as a function are used. In this work, we provide a theoretical analysis to compare and contrast the two major branches of Bayesi...
Conference Paper
Full-text available
Mapping the occupancy level of an environment is important for a robot to navigate in unknown and unstructured environments. To this end, continuous occupancy mapping techniques which express the probability of a location as a function are used. In this work, we provide a theoretical analysis to compare and contrast the two major branches of Bayesi...
Conference Paper
Full-text available
Hilbert mapping is an efficient technique for building continuous occupancy maps from depth sensors such as LiDAR in static environments. However, to make the map adaptable to dynamic environments, its parameters need to be learned automatically. In this paper, we take a variational Bayesian approach to this problem, thus eliminating the regulariza...
Conference Paper
Understanding the dynamics of urban environments is crucial for path planning and safe navigation. However, the dynamics might be extremely complex making learning the environment an unfathomable task. Within the methods available for learning dynamic environments, dynamic Gaussian process occupancy maps (DGPOM) are very attractive because they can...
Conference Paper
Full-text available
We consider the problem of building continuous occupancy representations in dynamic environments for robotics applications. The problem has hardly been discussed previously due to the complexity of patterns in urban environments, which have both spatial and temporal dependencies. We address the problem as learning a kernel classifier on an efficien...
Article
Full-text available
Use of touch-screen-based interactions is growing rapidly. Hence, knowing the maneuvering efficacy of touch screens relative to other pointing devices is of great importance in the context of graphical user interfaces. Movement time, accuracy, and user preferences of four pointing device settings were evaluated on a computer with 14 participants ag...
Article
Understanding and predicting how influenza propagates is vital to reduce its impact. In this paper we develop a nonparametric model based on Gaussian process (GP) regression to capture the complex spatial and temporal dependencies present in the data. A stochastic variational inference approach was adopted to address scalability. Rather than modeli...
Conference Paper
Full-text available
Understanding and predicting how influenza propagates is vital to reduce its impact. In this paper we develop a nonparametric model based on Gaussian process (GP) regression to capture the complex spatial and temporal dependencies present in the data. A stochastic variational inference approach was adopted to address scalability. Rather than modeli...
Conference Paper
Malaria is a globally widespread mosquito-borne disease which is caused by Plasmodium parasite. Plasmodium falciparum is the most ubiquitous among few species of Plasmodium and its Gametocyte stage is the most virulent among all stages and species. Although blood films are stained for better visualisation through the microscope, the color differenc...
Article
Targeting and tracking in graphical user interfaces have been widely studied, but attempts to model targeted-tracking are few. Targeted-tracking is essentially a two component task of tracking followed by targeting, where either or both components may dominate depending on the levels of difficulty in each component. The applicability of an empirica...
Article
Full-text available
Current models for targeted-tracking are discussed and shown to be inadequate as a means of understanding the combined task of tracking, as in the Drury's paradigm, and having a final target to be aimed at, as in the Fitts' paradigm. It is shown that the task has to be split into components that are, in general, performed sequentially and have a mo...
Article
There is a possibility that computer mice may be replaced with eye-gaze or touchscreen technologies. Hence, it is imperative that we investigate the effect of the type of input device in conditions having lateral constraints. A set of tracks with different levels of difficulty were tested. The type of device had little influence on movement time fo...
Conference Paper
A pointing device plays an important role in human-computer interaction. The computer mouse is a convenient device for both pointing and steering. The literature related to the effect of mouse gain on steering tasks is scarce. An experiment was conducted with 10 participants and each participant was asked to traverse a constrained path using a comp...
Conference Paper
Full-text available
The optical microscope is extensively used in a variety of biotechnological studies. However, most of the biologists and allied-medical technologists face utter inconvenience with their medical testing tasks and researches as they have to keep their eye on a microscope half of the day. Likewise, they have to focus the microscope using two mechanica...
Conference Paper
Adopting a hand gesture based appliance control system would be a virtuous idea for smart homes. However, complexity of the home background makes it challenging to work such systems in real home environments. In this paper, we present developing a robust system which can practically be used in complex backgrounds. The users are required to wear nei...

Network

Cited By

Projects

Projects (4)
Project
To delimitate important segments of images from the rest of the background by extracting features that are invariant to translation, rotation and scale.
Project
With the aim of developing safe robotic systems, my research focuses on developing machine learning algorithms for modeling spatiotemporal uncertainty in unstructured and dynamic environments. www.ransalu.com
Project
Considering open-loop and visual feedback control of hand/arm movements in combined targeting tracking tasks of the human-computer interface, a statistical model was developed. The model is a set of four equations derived based on statistical analysis of data collected from human subject based experiments. The model is important in: * Design and evaluation of input devices * Design and evaluation of graphical user interfaces (GUI) * Design and evaluation of virtual environments such as computer games Plausibly, the experimental results can be exploited to mimic hand movement dynamics in robotic manipulation. We developed a predictive model for targeted-tracking tasks. The model was experimentally verified for different input devices.