Our recent paper presenting a normative multi-objective based intersection collision avoidance system will be presented on KES in June 2022
Abstract – Coordination of vehicles’ behaviour is critical at intersections to avoid collisions. Intersection management is even more challenging when it is required to avoid collisions and address multiple objectives such as improving the traffic flow and decreasing vehicles’ travel time simultaneously in real-time. Normative Multi-agent systems enable the development of centralised collision avoidance techniques by synthesis-ing and applying traffic norms in real-time. However, they do not take into account the impact of the synthesised norms on other objectives that must be met by the system. In this paper, the proposed normative multi-objective based intersection collision avoidance system is based on a Hybrid Norms Synthesising (HNS) model that avoids collisions using run-time synthesised norms. HNS enables the vehicles to synthesise local norms, considering their local traffic flow, which results in decreasing their waiting time and creating a smoother traffic flow while avoiding the collisions at the same time. The simulation results show that HNS outperforms the state-of-the-art algorithm IRON in various scenarios.
Our paper on satisfying user preferences in optimised ridesharing services: a multi-agent multi-objective optimisation approach is published on 22 January 2022 in Applied Intelligence Journal
Abstract – Ridesharing services offer on-demand transportation solutions while improving the utilization of the available capacity of the vehicles on the road. Profit, travel time, and cost are the commonly optimized objectives in current ridesharing services. Existing multi-objective optimization-based services mainly focus on maximizing profit for service providers while minimizing the travel time for passengers. However, various personal preferences (e.g., co-riders’ gender for a passenger or preferred area of service for a driver) should be considered when offering such services. Such preferences are often conflicting with one another and with the objectives such as cost and travel time. Therefore finding an optimized solution and satisfying such preferences simultaneously is challenging. To address this challenge, this paper proposes a Multi-agent, Multi-objective, Preference-based ridesharing model (MaMoP) that offers an optimized ridesharing solution while satisfying users’ preferences simultaneously. MaMoP uses social-reasoning techniques to model preferences and their relations and employs evolutionary algorithms to find an optimized solution.
https://link.springer.com/article/10.1007/s10489-021-02887-1
Our paper Intelligent Shared Mobility Systems: A Survey on Whole System Design Requirements, Challenges and Future Direction is published in IEEE Access
Abstract – Shared Mobility Systems (SMS) facilitate on-demand journeys using one or more transportation modes such as car-sharing, bike-sharing, or ride-sharing. As a result, SMS often face challenges such as finding suitable facility locations, efficient routing of shared vehicles, matching and re-distributing available resources with dynamic demands. Most existing surveys study how a particular challenge is addressed using artificial intelligence, machine learning, and optimisation techniques. However, these surveys fail to address the crucial “Whole System Design” point of view, which includes the “whole system” of interconnected stakeholders, entities, and subsystems that participate in, impact, and influence the success of each other and system a whole. Such a survey is highly required with the growing demand for flexible SMS that supports autonomous decision-making and offers multi-modal and inter-operable transportation services catered for highly dynamic traffic conditions in urban areas. This paper attempts to fill this gap by categorising the SMS’ interconnected challenges in different transportation modes and reviewing how offered solutions across all modes address these challenges as a unified system.
Our recent paper Adaptive Workload Orchestration in Pure Edge Computing: A Reinforcement-Learning Model is published in ICTAI 2021
Abstract – Edge computing is a promising paradigm that can address the requirements of compute-intensive tasks generated by delay-sensitive applications, through bringing processing and storage to the edge of the network. Task offloading is challenging in open and dynamic environments where applications with various Service Level Agreement (SLA) and Quality of Service (QoS) requirements frequently produce a fluctuated workload at the edge of a network with heterogeneous, mobile, and geodistributed nodes.The current literature has addressed this challenge by offloading tasks to fog or Mobile Edge Computing (MEC) servers. However, in strictly delay-sensitive applications such as augmented reality, autonomous driving, or remote surgery, a Pure Edge Computing (PEC) paradigm that allows peer-to-peer communication and cooperation is more reasonable.This paper proposes a novel learning-based task offloading model that enables a pure edge-based system with mobile and resource-constrained nodes to accommodate fluctuating workload generated by applications with various SLAs and QoS. The results show a better utilization of resources and tasks success rate when compared to the state-of-the-art algorithms.
Our paper Urban Emergency Management using Intelligent Traffic Systems: Challenges and Future Directions is published in ISC2 2021
Abstract – Emergency management systems’ performance relies on effective utilisation of available resources and timely delivery of designated services. Intelligent transportation systems are key enablers of such systems in urban areas, where both resource utilisation and service delivery are hugely impacted by the accessibility of the road network. Using machine learning techniques, and accessing big data and the Internet of Things has already made tremendous advancements in the field of intelligent transportation systems and emergency management. However, the assumption of having access to historical data and predictive modelling is not always practical as emergency situations may occur unanticipated. In this paper, we briefly review the most recent related work, discuss the existing challenges and highlight future directions. We also present some preliminary results wherein the absence of historical data, ontological knowledge, and normative systems are used to improve the emergency service delivery and avoid new accidents while keeping the overall system performance. In this paper, we briefly review the most recent related work, discuss the existing challenges and highlight future directions. We also present some preliminary results wherein the absence of historical data, ontological knowledge, and normative systems are used to improve the emergency service delivery and avoid new accidents while keeping the overall system performance.
Our recent paper on using ontology to guide reinforcement learning agents in unseen situations has just been published in Applied Intelligence Journal.
Abstract – In multi-agent systems, goal achievement is challenging when agents operate in ever-changing environments and face unseen situations, where not all the goals are known or predefined. In such cases, agents need to identify the changes and adapt their behavior, by evolving their goals or even generating new goals to address the emerging requirements.
Learning and practical reasoning techniques have been used to enable agents with limited knowledge to adapt to new circumstances. However, they depend on the availability of large amounts of data, require long exploration periods, and cannot help agents to set new goals. Furthermore, the accuracy of agents’ actions is improved by introducing added intelligence through integrating conceptual features extracted from ontologies. However, the concerns related to taking suitable actions when unseen situations occur are not addressed.
This paper proposes a new Automatic Goal Generation Model (AGGM) that enables agents to create new goals to handle unseen situations and to adapt to their ever-changing environment on a real-time basis. AGGM is compared to Q-learning, SARSA, and Deep Q Network in a Traffic Signal Control System case study. The results show that AGGM outperforms the baseline algorithms in unseen situations while handling the seen situations as well as the baseline algorithms.
https://link.springer.com/article/10.1007/s10489-021-02449-5
Our new publication Runtime Norm Synthesis in Multi-Objective Multi-Agent Systems is accepted by COINE, co-located with AAMAS 2021
Abstract – Norms represent behavioural aspects that are encouraged by a social group of agents or the majority of agents in a system. Normative systems enable coordinating synthesised norms of heterogeneous agents in complex multi-agent systems autonomously. In real applications, agents have multiple objectives that may contradict each other or contradict the synthesised norms. Therefore, agents need a mechanism to understand the impact of a suggested norm on their objectives and decide whether or not to adopt it. To address these challenges, a utility based norm synthesis (UNS) model is proposed which allows the agents to coordinate their behaviour while achieving their conflicting objectives. UNS proposes a utility-based case-based reasoning technique, using case-based reasoning for run-time norm synthesising in a centralised approach, and a utility function derived from the objectives of the system and its operating agents to decide whether or not to adopt a norm. The model is evaluated using a two intersecting roads scenario and the results show its efficacy to optimise multiple objectives while adopting synthesised norms.
Our recent paper on an ontology-based intelligent traffic signal control model has just been accepted in IEEE International Intelligent Transportation Systems Conference.
Abstract – Reinforcement Learning (RL) can enhance the adjustment of the traffic signals’ phases to improve the traffic flow. RL methods use ontologies and reasoning to enrich the controllers’ domain knowledge, enabling them to interpret the traffic data, and ultimately improving their performance.
Various RL methods are proposed for signal controllers with assumptions such as operating in non-stochastic environments with a predictable traffic flow and observing the fine-grained information of all vehicles. Such methods have not examined the robustness of the trained RL controllers’ action selection when deployed in dynamic environments with partial detection of vehicles. However, in the real world, not all vehicles can be detectable, and not all events can be predicted.
In this paper, we propose an Ontology-based Intelligent Traffic Signal Control (OITSC) model that augments the RL controllers’ observation using an environment ontology model, which improves their action selection particularly in dynamic, partially observable environments with stochastic traffic flow. The decreased vehicles’ waiting time in various traffic scenarios with partial detection of vehicles, noisy sensor data, and unexpected traffic events shows that the performance of the controllers is significantly improved in all tested RL algorithms (i.e., Q-learning, SARSA, and Deep Q-Network).