Performance In Simulated Environments Research Articles

Target-driven multi-input mapless robot navigation with deep reinforcement learning

Collision-free navigation without pre-built maps is a promising and challenging problem. To make full use of the information in the environment, we propose a multi-input mapless navigation framework that combines both vision-based and ranging-based approaches. It takes the position and orientation of the robot relative to the target as the reference, incorporating RGB-D images and distance measurements as state inputs to the deep reinforcement learning model. In the framework, Dual-VAE is designed to encode RGB images and depth images simultaneously, in order to effectively mine the intrinsic information of RGB-D images. Moreover, we engineer a special reward function that enables the robot to navigate with a smooth trajectory and leave redundant space with obstacles for other tasks. Target-driven multi-input mapless navigation framework has been experimentally proven to achieve satisfactory performance in simulation environments. Our source code and simulation environments are released at https://github.com/YosephYu/MultiInput-Mapless-Navigation.

Evaluating the dual-task decrement within a simulated environment: Word recall and visual search

Simulated environments have become better able to replicate the real world and can be used for a variety of purposes, such as testing new technology without any of the costs or risks associated with working in the real world. Because of this, it is now possible to gain a better understanding of cognitive demands when working in operational environments, where individuals are often required to multitask. Multitasking often results in performance decrements, where adding more tasks can cause a decrease in performance in each of the individual tasks. However, little research investigated multitasking performance in simulated environments. In the current study we examined how multitasking affects performance in simulated environments. Forty-eight participants performed a dual visual search and word memory task where participants were navigated through a simulated environment while being presented with words. Performance was then compared to single-task performance (visual search and word memory alone). Results showed that participants experienced significant dual-task interference when comparing the dual-tasks to the single-tasks and subjective measures confirmed these findings. These results could provide useful insight for the design of technology in operational environments, but also serve as an evaluation of MRT in simulated environments.

Robust feedback control of ZMP-based gait for the humanoid robot Nao

Numerous approaches have been proposed to generate well-balanced gaits in biped robots that show excellent performance in simulated environments. However, in general, the dynamic balance of the robots decreases dramatically when these methods are tested in physical platforms. Since humanoid robots are intended to collaborate with humans and operate in everyday environments, it is of paramount importance to test such approaches both in physical platforms and under severe conditions. In this work, the special characteristics of the Nao humanoid platform are analyzed and a control system that allows robust walking and disturbance rejection is proposed. This approach combines the zero moment point (ZMP) stability criterion with angular momentum suppression and step timing control. The proposed method is especially suitable for platforms with limited computational resources and sensory and sensory-motor capabilities.

Objective Fidelity Evaluation in Multisensory Virtual Environments: Auditory Cue Fidelity in Flight Simulation

We argue that objective fidelity evaluation of virtual environments, such as flight simulation, should be human-performance-centred and task-specific rather than measure the match between simulation and physical reality. We show how principled experimental paradigms and behavioural models to quantify human performance in simulated environments that have emerged from research in multisensory perception provide a framework for the objective evaluation of the contribution of individual cues to human performance measures of fidelity. We present three examples in a flight simulation environment as a case study: Experiment 1: Detection and categorisation of auditory and kinematic motion cues; Experiment 2: Performance evaluation in a target-tracking task; Experiment 3: Transferrable learning of auditory motion cues. We show how the contribution of individual cues to human performance can be robustly evaluated for each task and that the contribution is highly task dependent. The same auditory cues that can be discriminated and are optimally integrated in experiment 1, do not contribute to target-tracking performance in an in-flight refuelling simulation without training, experiment 2. In experiment 3, however, we demonstrate that the auditory cue leads to significant, transferrable, performance improvements with training. We conclude that objective fidelity evaluation requires a task-specific analysis of the contribution of individual cues.

Overcoming psychological barriers to good discovery decisions

Better individual and team decision-making should enhance R&D performance. Reproducible biases affecting human decision-making, known as cognitive biases, are well understood by psychologists. These threaten objectivity and balance and so are credible causes for continuing unpleasant surprises in Development and high operating costs. For four of the most common and insidious cognitive biases, we consider the risks to R&D decision-making and contrast current practice with use of evidence-based medicine by healthcare practitioners. Feedback on problem-solving performance in simulated environments could be one of the simplest ways to help teams improve their selection of compounds and effective screening sequences. Computational tools that encourage objective consideration of all of the available information might also contribute.

On the Ability of Humans to Haptically Identify and Discriminate Real and Simulated Objects

The ability of human subjects to identify and discriminate between different-sized real objects was compared with their ability to identify and discriminate between different-sized simulated objects generated by a haptic interface. This comparison was additionally performed for cases of limited force and limited stiffness output from the haptic device, which in effect decrease the fidelity of the haptic simulation. Results indicate that performance of size-identification tasks with haptic-interface hardware capable of a minimum of 3 N of maximum force output can approach performance in real environments, but falls short when virtual surface stiffness is limited. For size-discrimination tasks, performance in simulated environments was consistently lower than performance in a comparable real environment. Interestingly, significant variations in the fidelity of the haptic simulation do not appear to significantly alter the ability of a subject to identify or discriminate between the types of simulated objects described herein.