174 Table 2: Continued. Author (s) Key concepts Li et al. (2002) A hierarchical control scheme is developed to enable multiple Unmanned Combat Air Vehicles (UCAVs) autonomously achieving demanding missions in hostile environments. The scheme consists of four layers: 1) a high-level path planner, 2) a low-level path planner, 3) a trajectory generator and 4) a formation control algorithm. The high-level path planner plans a path based on a Voronoi diagram that compromises the cost involved in exposing to the threats and the costs of fuel expense based on static threat and target information provided by the command centre. The low-level path planner plans a finer grained path from the waypoint provided by the high-level planner to the current position. It also checks if there is a popup threat on the route and will plan a path to avoid it while reaching the next waypoint. The trajectory generator computes the control input for the leader based on feasible trajectories for the UCAVs to follow. The leader’s position and input are transferred to the formation controller that assures that each follower will maintain formation regardless of the manoeuvres of the leader. Lindner et al. (2017) HERA (Hybrid Ethical Reasoning Agents) is a software library to model autonomous moral decision-making. It represents the robot’s possible actions together with the causal chains of consequences the actions initiate. Logical formulae are used to model ethical principles. HERA implements several ethical principles, such as a Pareto-inspired principle, the principle of Double Effect and utilitarianism. The applied format is called a causal agency model. It reduces determining moral permissibility by checking if principle-specific logical formulae are satisfied in a causal agency model. Ricard and Kolitz (2003) The ADEPT (All-Domain Execution and Planning Technology) architecture for intelligent autonomy is a hierarchical extension of the sense-think-act paradigm of intelligence and is closely related to the Observe-Orient-Decide-Act (OODA) loop that is often used in the military. Intelligent autonomy has to deal with 3 challenges. Firstly, the executed plans and activities have to meet mission objectives and abide the constraints. Secondly, it has to cope with uncertainties and thirdly, it has to real-time dynamically adjust a vehicle’s plan due to changes in context and environment. ADEPT is a reusable object-oriented software framework that consists of four modules: 1) situation assessment, 2) plan generation, 3) plan implementation and 4) coordination. Vanderelst and Winfield (2018) Ethical behaviour in robots is implemented by simulation theory of cognition in which internal simulations for actions and prediction of consequences are used to make ethical decisions. The method is a form of robot imagery and does not make use of verification of logical statements that is often used to check if actions are in accordance with ethical principles. It is an additional or substitute framework for implementing robotic ethics as alternative for logic-based AI that currently dominates the field. The method uses a separate Ethical layer that is independent of the robot’s controller. The robot controller generates a set of behavioural alternatives, the Simulation Module simulates the consequences of each alternative which are in turn evaluated by the Evaluation Module and calculated in a single metric that reflects the desirability of a certain action. The output of the evaluation of each alternative is sent to the robot controller. The Simulation model consists of three components, a model of the robot controller, a model of the domain specific human and a world model. This allows for the ethics of higher-level goals to be evaluated based on the actions that they induce and the prediction of ultimate consequences of tasks and goals. APPENDIX C
RkJQdWJsaXNoZXIy MjY0ODMw