Laura Humphrey's research while affiliated with Air Force Research Laboratory and other places

... However, these abstract statuses are usually difficult to quantify. Many studies have applied self-reported surveys for assessing different aspects of the cognitive status, such as the National Aeronautics and Space Administration Task Load Index (NASA-TLX) [13] and the Rating Scale Mental Effort (RSME) [26] for assessing mental workload, the Dundee Stress State Questionnaire (DSSQ) [27] for assessing task engagement and distraction, the Karolinska Sleepiness Scale (KSS) [28] for assessing fatigue and drowsiness, and the Situation Awareness Rating Technique (SART) [29] for assessing situational awareness. These instruments have been adopted and validated in many studies and provide a solution to quantify cognitive statuses. ...

... Although model trustworthiness is an on-going area of research in artificial intelligence and computer science literature (Chang et al., 2005;Wei et al., 2020;Markus et al., 2021), trust was not measured at any point in this thesis. Measuring the trustworthiness of a model has traditionally been done through surveys, resulting in psychometric measures of perceived trust (Miller, 2022). ...

... This inherently makes trust measurement and modeling controversial. There is an ongoing debate about the validity of psychometric measures: whether they are real or merely artificial/folk constructs (Annett, 2002;Dekker & Hollnagel, 2004;Dekker & Nyce, 2015); whether they converge with objective measures (Dekker & Nyce, 2015;Matthews et al., 2020); and whether they satisfy the requirements for cardinal levels of measurement (Bolton et al. n.d.-a, n.d.-b;Wei et al., 2019Wei et al., , 2020. ...

... These proposed ideas cover a wide range of possibilities such as allowing UAVs to land at vertistops or vertiports installed on roofs of existing buildings or within cloverleaf exchanges on freeways. Several simulation tools [9]- [11] have also been developed. ...

... Linear temporal logic (LTL) is a well known notation for specifying concurrent behavior [19]. It has found several applications to robotics, including in motion planning [25] and multi-robot coordination [2], [4]. (These use either the GR(1) fragment or finite LTL, which have good synthesis algorithms.) ...

... In [9] the research discusses a method to perform run-time assurance for learning systems with an assurance architecture designed in Architecture Analysis and Design Language (AADL) and formal contracts for each of the components modeled and verified in Assume Guarantee (AGREE) annex. In [12] Davis discusses an approach to use architectural analysis to prove that the protocol designed for multi-agents satisfies the specified properties. This effort also emphasized the use of AADL and AGREE for formal assurance. ...

... To simplify matters, assume that the airplane (P) is tracking the movement of one vehicle (V) that is about to cross the runway. Let us further assume that P tracks V using a perception module that can only determine the position of the vehicle with a certain accuracy [33], i.e., for every position of V, the perception module reports a noisy variant of the position of V. However, it is important to know that the plane obtains a sequence of these measurements. ...

... For example, human trust has been investigated in a multiagent setting in several studies, e.g. (Williams et al., 2021;Wang et al., 2018). However, there is still work to be done in understanding how trust is affected when robots differ in levels of autonomy, and exhibit varied performance and failure rates. ...

... However, The k-stabilizing shield synthesis problem is unrealizable for many safety-critical systems, because a finite number of deviations cannot be guaranteed. Humphrey et al. [34] addressed this problem by proposing the notion of admissible shields which was extended and generalized in Könighofer et al. [35] by assuming that systems have a cooperative behavior with respect to the shield, i.e., the shield ensures a finite number of deviations if the system chooses certain outputs. The authors presented a synthesis procedure that maximizes the cooperation between system and environment for satisfying the required enforced properties. ...

... Consequently, robot systems are still in their infancy in developing the ability to explain their own behavior when confronting noisy sensory inputs and executing complex multistep decision processes. Planner-based robot systems can generally provide an interpretable account for their actions to humans [e.g., by Markov decision processes (7,8), HTN (9), or STRIPS (10)], but these planners struggle to explain how their symbolic-level knowledge is derived from low-level sensory inputs. In contrast, robots equipped with deep neural networks (DNNs) (11) have demonstrated impressive performance in certain specific tasks due to their powerful ability to handle low-level noisy sensory inputs (12,13). ...