Rob Duffield’s research while affiliated with University of Technology Sydney and other places

The key to robot-assisted rehabilitation lies in the design of the human-machine interface, which must accommodate the needs of both patients and machines. Current interface designs primarily focus on machine control algorithms, often requiring patients to spend considerable time adapting. In this paper, we introduce a novel approach based on the Cooperative Adaptive Markov Decision Process (CAMDPs) model to address the fundamental aspects of the interactive learning process, offering theoretical insights and practical guidance. We establish sufficient conditions for the convergence of CAMDPs and ensure the uniqueness of Nash equilibrium points. Leveraging these conditions, we guarantee the system's convergence to a unique Nash equilibrium point. Furthermore, we explore scenarios with multiple Nash equilibrium points, devising strategies to adjust both Value Evaluation and Policy Improvement algorithms to enhance the likelihood of converging to the global minimal Nash equilibrium point. Through numerical experiments, we illustrate the effectiveness of the proposed conditions and algorithms, demonstrating their applicability and robustness in practical settings. The proposed conditions for convergence and the identification of a unique optimal Nash equilibrium contribute to the development of more effective adaptive systems for human users in robot-assisted rehabilitation.

Purpose: This study investigated the relationship between travel demands and match loads on perceptual recovery, fatigue, and sleep following postmatch travel in national football teams. In addition, the influence of travel demands and the time between match kickoff and travel departure on postmatch recovery was examined. Methods: Match-running load (via GPS) and travel data were obtained from 79 male national-team footballers. Postmatch travel duration, direction, context, time-zone difference, and time between kickoff and travel departure were collated. Athletes provided perceptual ratings of fatigue, soreness, sleep, stress, and recovery from 1 day prematch through to 3 days after postmatch travel. Linear mixed models assessed the influence of match load and travel on perceptual ratings for 3 days posttravel. Additional models assessed a standardized postmatch time point of match day + 3 to determine whether timing and extent of travel influenced recovery. Results: Higher match loads were associated with poorer recovery, fatigue, soreness, and sleep (P = .001-.032). Athletes reported poorer fatigue, soreness, and recovery when traveling from national teams back to clubs compared with between national-team matches (P < .001). Traveling eastward was associated with poorer sleep (P = .004). Longer periods between kickoff and travel departure were associated with poorer sleep on match day + 3 (P = .003). Conclusions: Perceptual recovery, fatigue, sleep, and soreness following postmatch travel were affected by both match load and travel demands. Greater match loads and eastward postmatch travel may impair recovery. In addition, departing later after a match was associated with poorer sleep on match day + 3; however, no relationships existed to any other recovery measures.

Purpose: To explore the influence of menstrual phase on postmatch perceptual responses and the time course of recovery for professional footballers. Methods: Thirteen naturally menstruating footballers tracked their menstrual cycle and reported perceptual responses for up to 4 cycles. Menstrual phases were determined by calendar-based tracking and urinary hormone tests and classified as menstruation, follicular, or luteal. On match days (MDs) and the following 2 days (MD+1 and MD+2), players completed perceptual questionnaires on fatigue, soreness, stress, sleep, and perceived recovery status (PRS). Total high-speed running distance during matches was recorded using GPS devices to represent load. Data were analyzed using linear mixed models. Results: Day × total high-speed running distance affected PRS (P < .001), total wellness (P < .001), fatigue (P = .047), soreness (P < .001), and stress (P = .044). Significant main effects were found for menstrual phase on PRS (P = .038), Day on stress (P = .034), and total high-speed running distance on soreness (P = .045). During the menstruation phase, moderate effect sizes (ES) existed for worse PRS on MD and MD+2 (P = .07-.28, ES > 0.51) and better sleep quality on MD+2 (P = .13, ES = 0.56). No significant differences between menstrual phases existed (P > .05), and all other ESs were trivial to small (ES < 0.50). All perceptual measures, except stress, differed significantly between days (P < .05), with differences based on match load and, to a lesser extent, menstrual phase. Conclusion: PRS may be worse during menstruation, although menstrual phase has limited association with postmatch perceptual responses. Variability in the recovery time course for perceptual measures exists between menstrual phases, but evidence for consistently impaired recovery time course in any phase was not evident.

This study investigated the association between environmental temperature and match‐play characteristics (shooting, passing, dribbling and defending) in four professional football leagues. Twenty‐seven performance indicators (PI's) were collated from 1585 matches from the German Bundesliga 1 and 2, Spanish La Liga and Australian A‐League. Environmental data were obtained for dry‐bulb temperature (T) and wet‐bulb globe temperature (WBGT) retrospectively from public sources. For each league, linear regressions were used to determine relationships between PI's and T and WBGT and linear mixed models were used to determine those associations across all four leagues. Individual leagues showed varying associations between a collection of PI's and environmental measures. When combining the four leagues' match data, 8 of the 17 investigated parameters were associated with T and WBGT (p < 0.002). Passes, especially short passes, were reduced in higher T (−2.3 [−3.1 to −1.5] and p < 0.001) and WBGT (−3.1 [−4.0 to −2.1] and p < 0.001), alongside an increase in the success rate of passes (0.06 [0.02–0.09] and p ≤ 0.001). The number of passes into the opponent's final third was reduced for both T (−0.18 [−0.25 to −0.05] and p = 0.001) and WBGT (−0.17 [−0.28 to−0.05] and p = 0.002), but the number of key passes leading to a shot or goal was not associated with T or WBGT (p ≥ 0.67). The number of touches, take‐ons and turnovers were reduced in higher T and WBGT (all p < 0.001). Accordingly, in higher heat stress, match actions, especially those performed at high volumes, are reduced. Therefore, teams should expect a possibly altered match play and may consider adapting tactical or heat‐mitigating strategies to counter these effects.

A cross-sectional analysis was performed to investigate associations between environmental temperatures and injury occurrence in two professional male football (soccer) leagues. Data from seven seasons of the German Bundesliga (2142 matches) and four seasons of the Australian A-League (470 matches) were included. Injuries were collated via media reports for the Bundesliga and via team staff reports in the A-League and comprised injury incidence, mechanisms (contact, noncontact), locations (e.g., ankle, knee, and thigh), and types (e.g., muscle and tendon, joint and ligament). Weather data included ambient air temperature (temperature or T) and wet bulb globe temperature (WBGT), which were collected from online sources retrospectively. Generalized linear mixed models were analyzed to examine associations between temperature or WBGT and injury occurrence for each league, respectively. Additionally, matches were grouped into categories of 5°C temperature steps to compare for injury occurrence. Results showed no relationship existed between either temperature or WBGT and any injury occurrence, mechanisms, locations or types for the Bundesliga ( P > 0.10). A trend for an increase in injury occurrence in higher WBGT existed in the A-League ( P = 0.05). Comparisons between 5°C temperature categories showed no significant differences for injury occurrence for either temperature or WBGT in either League ( P > 0.05). Within the observed temperature ranges (−11.2 to 37.1°C T; −12.2 to 29.6°C WBGT) environmental temperature had no relationship with the rate or type of injury occurrence in professional football. Nevertheless, the number of matches at extreme heat within this study was limited and other factors (e.g., playing intensity, season stage, ground conditions) likely co-influence the relationship with injuries.

This study investigated the association of menstrual cycle phase and symptoms with objective and subjective sleep measures from professional footballers before and after matches. Twenty‐three non‐hormonal contraceptive‐using professional footballers (from four clubs) were monitored for up to four menstrual cycles during a domestic league season. Menstrual phases (menstruation, mid‐late follicular, luteal) were determined using calendar counting and urinary hormone tests (luteinizing hormone and pregnandiol‐3‐glucuronide). Players rated the severity of 18 symptoms on the evenings of matches and the following two evenings. Individual daily summed menstrual symptom severity (MSS) scores were calculated. Subjective sleep quality was rated the morning of matches and the following two mornings. Objective sleep (bedtime, waketime, total sleep time [TST], sleep onset latency, wake after sleep onset, sleep efficiency) obtained from actigraphy was measured the night prior to (MN−1), of (MN), and following (MN+1) matches. Linear mixed models were performed for each sleep measure to examine the effects of menstrual phase and symptoms. Bedtime was significantly later for MN (p < 0.001), waketime was significantly earlier for MN+1 (p < 0.001), and TST was significantly longer for MN−1 (p < 0.001). Menstrual phase did not have a significant effect on any sleep variable (p > 0.05). Increased MSS score was associated with increased TST (p = 0.03) and increased waketime (p = 0.03). Increased lower back pain severity and mood changes/anxiety severity were associated with increased waketime (p = 0.048) and TST (p = 0.009), respectively. Overall, bedtime and waketime were affected by the night related to matches, with increased TST the night before a match. Menstrual phase was not related to any objective or subjective sleep variables, whilst increasing menstrual symptom severity was related to later waketime and longer TST.

Purpose : This study investigated associations between environmental temperatures and match running performance in 4 professional football leagues. Methods : Running performance indicators including total, high-speed, and sprint distances were collated from 1610 matches from the German Bundesliga 1 and 2, Japanese J-League, and Turkish SüperLig. Environmental data for each of these matches were obtained for dry-bulb and wet-bulb globe temperatures (WBGT) retrospectively from public sources. Linear regressions were used to determine relationships between running performance indicators and both temperature and WBGT for individual leagues. Furthermore, linear mixed models were used to determine associations across all 4 leagues, accounting for differences between them as random effects. Bonferroni corrections were applied to account for multiple tests. Results : Overall, combined-league data showed that total distance (95% CI, −0.50 to 0.37; β : −0.36), number of high-speed runs (95% CI, −4.57 to 2.93; β : −0.29), high-speed distances (95% CI, −0.07 to 0.05; β : −0.28), number of sprints (95% CI, −2.72 to 2.07; β : −0.39), and sprint distances (95% CI, −0.05 to 0.03; β : −0.22) were all lower when WBGT was higher ( P < .001), whereas the peak speed recorded per match (95% CI, 0.01 to 0.03; β : 0.18) was higher when WBGT was higher ( P < .001). Models with temperature instead of WBGT derived similar results. Conclusion : Warmer environmental conditions were associated with lower total, high-speed, and sprint distances covered. These responses may result from an increased thermoregulatory load or indirectly from an adapted individual or team-tactical pacing strategy in warmer conditions. Teams should consider strategies to counter such effects to avoid lower distances covered at high intensities that are related to success in football.

Purpose: This study examined the relationship of prematch travel and in-camp training on ensuing physical and technical match performance of footballers (soccer) competing for a national team. Methods: Match running and technical performance data were obtained from 68 national-team footballers competing in international matches (N = 108). Match performance data were aligned with the confirmed travel durations, time-zone change, travel direction, and time between arrival and kickoff for the travel to the match. In addition, in-camp training load from the 3 days prior to national-team matches was also collated. Linear mixed models assessed relationships between travel and training measures with physical and technical match performance outcomes. Results: Travel variables explained little variance in outcomes (R2 = .02-.16). Traveling eastward was associated with an increase in total (P = .042) and very high-speed distance (P = .030) in matches and a 5% decrease in pass accuracy (P = .012). Greater time-zone difference was associated with increased match decelerations (P = .027), while arriving later for matches was associated with increased total (P = .041) tackles. Increases in training high-speed distance were associated with increases in match high-speed distance (P = .004) and number of decelerations (P < .001). An increase in training total distance was associated with a decrease in match decelerations (P = .002). Conclusions: Prematch travel appeared to have minimal effects on physical and technical performance in this national team. Training loads prior to matches may have some relationship with match running performance, and, thus, readiness to train should be a priority for athletes arriving in national-team camps.