Search

A novel deep multi-agent reinforcement learning framework is proposed to identify and resolve conflicts among a variable number of aircraft in a high-density, stochastic, and dynamic en route secto...

The use of electrical vertical takeoff and landing (eVTOL) aircraft to provide efficient, high-speed, on-demand air transportation within a metropolitan area is a topic of increasing interest, which is expected to bring fundamental changes to the city infrastructures and daily commutes. NASA, Uber, and Airbus have been exploring this exciting concept of Urban Air Mobility (UAM), which has the potential to provide meaningful door-to-door trip time savings compared with automobiles. However, the ability to manage many of these eVTOL aircraft safely in a congested urban area presents a challenge unprecedented in air traffic management. In order to enable safe and efficient autonomous on-demand free flight operations in UAM, a computational guidance algorithm with collision avoidance capability is designed and analyzed. The approach proposed in this paper is to formulate this problem as a Markov Decision Process (MDP) and solve it using an online algorithm Monte Carlo Tree Search (MCTS). For illustration, a high-density free flight airspace simulator is created to test the proposed algorithm’s performance. Numerical experiment results show that this proposed algorithm has fewer conflicts and near mid-air collisions than Optimal Reciprocal Collision Avoidance (ORCA), a state-of-the-art collision avoidance strategy.

Electric vertical takeoff and landing vehicles are becoming promising for on-demand air transportation in urban air mobility (UAM). However, successfully bringing such vehicles and airspace operati...

An approach is given, which tries to design a robust controller using the plicy iteration algorithm for the nonlinear systems due to uncertainties. The controller design problem is divided into two steps. The first step is introducting the uncertainties into the cost function which changes the robust difficult point into the optimal problem. Second, to obtain the control strategy for the optimization of the new cosst function, the policy iteration technique is utilized to deal with the Hamilton-Jacobi-Bellman (HJB) equation. In the implementation of the given approach, the cost function is obtained approximately by constructing the critic neural networks while the optimal control strategy is derived directly based on system dynamic the critic neural network. Furthermore, the convergence for the uncertain system and the equivalence of the solution to the two kind of questions is proved. Final, the numerical results illustrated the correct of the presented method.

Obstacle avoidance for small unmanned aircraft is vital for the safety of future urban air mobility (UAM) and Unmanned Aircraft System (UAS) Traffic Management (UTM). There are many techniques for real-time robust drone guidance, but many of them solve in discretized airspace and control, which would require an additional path smoothing step to provide flexible commands for UAS. To provide a safe and efficient computational guidance of operations for unmanned aircraft, we explore the use of a deep reinforcement learning algorithm based on Proximal Policy Optimization (PPO) to guide autonomous UAS to their destinations while avoiding obstacles through continuous control. The proposed scenario state representation and reward function can map the continuous state space to continuous control for both heading angle and speed. To verify the performance of the proposed learning framework, we conducted numerical experiments with static and moving obstacles. Uncertainties associated with the environments and safety operation bounds are investigated in detail. Results show that the proposed model can provide accurate and robust guidance and resolve conflict with a success rate of over 99%.

To study the relationship between carbon dioxide combining power(CO₂-CP) and contrast-induced acute kidney injury (CI-AKI) in patients with ST segment elevation myocardial infarction and undergoing percutaneous coronary intervention.We retrospectively analyzed 174 patients admitted to our hospital from March 2012 to August 2013 with ST segment elevation myocardial infarction and underwent emergency percutaneous coronary intervention. Patients were divided into three tertiles according to pre-operative CO₂-CP: T1 (CO₂-CP22.62 mmol/L), T2(CO₂-CP 22.62-24.30 mmol/L), T3(CO₂-CP24.30 mmol/L). Baseline clinical data, CI-AKI incidence, in-hospital mortality and dialysis rate were compared among groups. An increase in serum creatinine of26.4 µmol/L and/or50% from baseline within 48 hours after contrast exposure was defined as CI-AKI. Univariate logistic regression analysis was used to identify the risk factors of CI-AKI. The relationship between CO₂-CP and CI-AKI was assessed by multivariate logistic regression analysis. Receiver operating characteristic curve was used to identify the optimal cutoff of the CO₂-CP for predicting CI-AKI.CI-AKI occurred in 25 (14.4%) patients, and lower CO₂-CP was related to higher incidence of CI-AKI (27.6% (16/58) in group T1, 5.3% (3/57) in group T2, 1.7 % (1/59) in group T3, P = 0.002) and higher in-hospital mortality (10.3% (6/58) vs. 0 and 1.7% (1/59), P = 0.010). Dialysis rate was similar among 3 groups (5.2% (3/58) vs. 0 and 1.7% (1/59), P = 0.168). The incidence of CI-AKI was significantly associated with CO₂-CP22.00 mmol/L in univariate analyses (OR = 6.767, 95% CI 2.731-16.768, P0.001). After adjusting for potential confounding risk factors, CO₂-CP22.00 mmol/L remained significantly associated with the incidence of CI-AKI (OR = 5.835, 95%CI 1.800-18.914, P = 0.003) in multivariate logistic regression. ROC analysis revealed that the optimal cutoff of CO₂-CP to predict CI-AKI was 22.00 mmol/L (sensitivity 64.0%, specificity 79.1%, AUC = 0.714).Pre-percutaneous coronary intervention CO₂-CP in patients with ST segment elevation myocardial infarction undergoing percutaneous coronary intervention is related to CI-AKI. CO₂-CP22.00 mmol/L predicts higher risk of CI-AKI in this patient cohort.