Sim-to-Real: A Performance Comparison of PPO, TD3, and SAC Reinforcement Learning Algorithms for Quadruped Walking Gait Generation

The performance of the state-of-the-art Deep Reinforcement algorithms such as Proximal Policy Optimization, Twin Delayed Deep Deterministic Policy Gradient, and Soft Actor-Critic for generating a quadruped walking gait in a virtual environment was presented in previous research work titled “A Comparison of PPO, TD3, and SAC Reinforcement Algorithms for Quadruped Walking Gait Generation”. We demonstrated that the Soft Actor-Critic Reinforcement algorithm had the best performance generating the walking gait for a quadruped in certain instances of sensor configurations in the virtual environment. In this work, we present the performance analysis of the state-of-the-art Deep Reinforcement algorithms above for quadruped walking gait generation in a physical environment. The performance is determined in the physical environment by transfer learning augmented by real-time reinforcement learning for gait generation on a physical quadruped. The performance is analyzed on a quadruped equipped with a range of sensors such as position tracking using a stereo camera, contact sensing of each of the robot legs through force resistive sensors, and proprioceptive information of the robot body and legs using nine inertial measurement units. The performance comparison is presented using the metrics associated with the walking gait: average forward velocity (m/s), average forward velocity variance, average lateral velocity (m/s), average lateral velocity variance, and quaternion root mean square deviation. The strengths and weaknesses of each algorithm for the given task on the physical quadruped are discussed.

A novel explosion-proof walking system: Twin dual-motor drive tracked units for coal mine rescue robots

A new explosion-proof walking system was designed for the coal mine rescue robot(CMRR) by optimizing the mechanical structure and control algorithm. The mechanical structure innovation lies mainly in the dual-motor drive tracked unit used, which showed high dynamic performance compared with the conventional tracked unit. The control algorithm, developed based on decision trees and neural networking, facilitates autonomous switching between "Velocity-driven Mode" and "Torquedriven Mode". To verify the feasibility and effectiveness of the control strategy, we built a self-designed test platform and used it to debug the control program; we then made a robot prototype and conducted further experiments on single-step, ramp, and rubble terrains. The results show that the proposed walking system has excellent dynamic performance and the control strategy is very efficient, suggesting that a robot with this type of explosion-proof walking system can be successfully applied in Chinese coal mines.

The Fuzzy Neural Network Control Scheme With H∞ Tracking Characteristic of Space Robot System With Dual-arm After Capturing a Spin Spacecraft

In this paper,the dynamic evolution for a dualarm space robot capturing a spacecraft is studied,the impact effect and the coordinated stabilization control problem for postimpact closed chain system are discussed.At first,the pre-impact dynamic equations of open chain dual-arm space robot are established by Lagrangian approach,and the dynamic equations of a spacecraft are obtained by Newton-Euler method.Based on the results,with the process of integral and simplify,the response of the dual-arm space robot impacted by the spacecraft is analyzed by momentum conservation law and force transfer law.The closed chain system is formed in the post-impact phase.Closed chain constraint equations are obtained by the constraints of closed-loop geometry and kinematics.With the closed chain constraint equations,the composite system dynamic equations are derived.Secondly,the recurrent fuzzy neural network control scheme is designed for calm motion of unstable closed chain system with uncertain system parameter.In order to overcome the effects of uncertain system inertial parameters,the recurrent fuzzy neural network is used to approximate the unknown part,the control method with H∞tracking characteristic.According to the Lyapunov theory,the global stability is demonstrated.Meanwhile,the weighted minimum-norm theory is introduced to distribute torques guarantee that cooperative operation between manipulators.At last,numerical examples simulate the response of the collision,and the efficiency of the control scheme is verified by the simulation results.

The Best Walking Recommendation Path by Considering the Quality of Street Space

Walking is a healthy and green way of travel,and seeking the best recommended route is the first choice for residents to travel.In this paper,according to actual demand of pedestrians,convenience,safety,comfort and appreciation are taken as different walking needs,to construct the evaluation system of urban street space quality.Combining the shortest path algorithm,the modified walking time cost of road section and intersection delay are used as impedance,obtaining the best walking recommendation paths of convenience,safety,comfort and appreciation between the beginning and the end.Moreover,the form of questionnaire survey is used for experimental verification of final recommended path.This paper could provide residents with walking path planning scheme under different needs.When guaranteeing travel efficiency,it could effectively improve walking quality of residents,with important significance for improving the attraction of green travel mode,providing diverse selectivity,and improving residents’ health.

Performance Analysis of Curved Track G2T-FSO(Ground-to-Train Free Space Optical)Model under Various Weather Conditions

The demand for broadband data services on high-speed trains is rapidly growing as more people commute between their homes and workplaces.However,current radio frequency(RF)technology cannot adequately meet this demand.In order to address the bandwidth constraint,a technique known as free space optics(FSO)has been proposed.This paper presents a mathematical derivation and formulation of curve track G2T-FSO(Ground-to-train Free Space Optical)model,where the track radius characteristics is 2667 m,divergence angle track is 1.5°for train velocity at V=250 km/h.Multiple transmitter configurations are proposed to maximize coverage range and enhance curve track G2T-FSO link performance under varying weather conditions.The curved track G2T-FSO model was evaluated in terms of received power,signal-to-noise ratio(SNR),bit error rate(BER),and eye diagrams.The results showed maximum coverage lengths of 618,505,365,and 240 m for 4Tx/1Rx,3Tx/1Rx,2Tx/1Rx,and 1Tx/1Rx configurations,respectively.The analyzed results demonstrate that the G2T-FSO link can be effectively implemented under various weather conditions.

Modeling Impacts on Space Situational Awareness PHD Filter Tracking

In recent years,probabilistic tracking methods have been becoming increasingly popular for solving the multi-target tracking problem in Space Situational Awareness(SSA).Bayesian frameworks have been used to describe the objects’of interest states and cardinality as point processes.The inputs of the Bayesian framework filters are a probabilistic description of the scene at hand,the probability of clutter during the observation,the probability of detection of the objects,the probability of object survival and birth rates,and in the state update,the measurement uncertainty and process noise for the propagation.However,in the filter derivation,the assumptions of Poisson distributions of the object prior and the clutter model are made.Extracting the first-order moments of the full Bayesian framework leads to a so-called Probability Hypothesis Density(PHD)filter.The first moment extraction of the PHD filter process is extremely sensitive to both the input parameters and the measurements.The specifics of the SSA problem and its probabilistic description are illustrated in this paper and compared to the assumptions that the PHD filter is based on.As an example,this paper shows the response of a Cardinality only PHD filter(only the number of objects is estimated,not their corresponding states)to different input parameterizations.The very simple Cardinality only PHD filter is chosen in order to clearly show the sole effects of the model mismatch that might be blurred with state estimation effects,such as non-linearity in the dynamical model,in a full PHD filter implementation.The simulated multi-target tracking scenario entails the observation of attitude stable and unstable geostationary objects.

A CLIMATOLOGY OF EXTRATROPICAL TRANSITION OF TROPICAL CYCLONES IN THE WESTERN NORTH PACIFIC

Based on best-track data and JRA-25 reanalysis,a climatology of western North Pacific extratropical transition (ET) of tropical cyclone (TC) is presented in this paper. It was found that 35% (318 out of 912) of all TCs underwent ET during 1979-2008. The warm-season (June through September) ETs account for 64% of all ET events with the most occurrence in September. The area 120°E-150°E and 20°N-40°N is the most favorable region for ET onsets in western North Pacific. The TCs experiencing ET at latitudes 30°N-40°N have the greatest intensity in contrast to those at other latitude bands. The distribution of ET onset locations shows obviously meridional migration in different seasons. A cyclone phase space (CPS) method was used to analyze the TC evolution during ET. Except for some cases of abnormal ET at relatively high latitudes,typical phase evolution paths-along which TC firstly showed thermal asymmetry and an upper-level cold core and then lost its low-level warm core-can be used to describe the main features of ET processes in western North Pacific. Some seasonal variations of ET evolution paths in CPS were also found at low latitudes south of 15°N,which suggests different ET onset mechanisms there. Further composite analysis concluded that warm-season ETs have generally two types of evolutions,but only one type in cold season (October through next May). The first type of warm-season ETs has less baroclinicity due to long distance between the TC and upper-level mid-latitude system. However,significant interactions between a mid-latitude upper-level trough and TC,which either approaches or is absorbed into the trough,and TC's relations with downstream and upstream upper-level jets,are the fingerprints for both a second type of warm-season ETs and almost all the cold-season ETs. For each type of ETs,detailed structural characteristics as well as precipitation distribution are illustrated by latitude.

Model-free adaptive control of space manipulator under different gravity environment

Due to the release of gravity in the space environment, the dynamic characteristics of the space manipulator have changed compared with that of the ground, which results in the change of its tracking precision. This paper presents a model-free adaptive control(MFAC) strategy to track the desired trajectory under different gravity environment. A dynamic transformation method and full form dynamic linearization(FFDL) approach are selected to dynamicly linearize the system, which can better eliminate the complex dynamics that may exist in the original system. The controlled object uses the two degrees of freedom of space manipulator and the controller only depends on the desired angle and torque of each joint of the space manipulator. Moreover, the proof of stability is also provided. Finally, simulation results are presented to demonstrate the effectiveness of the proposed strategy. It is shown that the proposed approach can achieve better trajectory tracking performance under different gravity environment without changing the control parameters, and the tracking precision can be significantly improved as compared with the proportional differential(PD) control results.

Hypersonic sliding target tracking in near space

To improve the tracking accuracy of hypersonic sliding target in near space,the influence of target hypersonic movement on radar detection and tracking is analyzed,and an IMM tracking algorithm is proposed based on radial velocity compensating and cancellation processing of high dynamic biases under the earth centered earth fixed(ECEF) coordinate.Based on the analysis of effect of target hypersonic movement,a measurement model is constructed to reduce the filter divergence which is caused by the model mismatch.The high dynamic biases due to the target hypersonic movement are approximately compensated through radial velocity estimation to achieve the hypersonic target tracking at low systematic biases in near space.The high dynamic biases are further eliminated by the cancellation processing of different radars,in which the track association problem can be solved when the dynamic biases are low.An IMM algorithm based on constant acceleration(CA),constant turning(CT) and Singer models is used to achieve the hypersonic sliding target tracking in near space.Simulation results show that the target tracking in near space can be achieved more effectively by using the proposed algorithm.

Sustainable living neighbourhoods:Measuring public space quality and walking environment in Lisbon

Obesity and global warming are two major challengesthat affect communities’health and quality of life.Providing good walking environments,with appropriate pedestrian infrastructure,green areas and access to commercial retail and transport hubs favour liveable,healthy,and sustainable cities.The Lisbon City Hall has implemented an intervention program in the public space as a strategy to improve environmental quality and mitigate the effects of climate change.The program“A square in each neighbourhood”intends not only to reinforce the economic base of Lisbon,namely tourism,restaurants,commerce,and creative activities but also to promote physical activity and improve the quality of living of all citizens.For studying the impact of the public space quality on the walking environment,Campolide,a typical Lisbon neighbourhood,Portugal,which includes one of the requalified squares,was selected.A two-step methodology based on spatial modelling is proposed.In the first step,the public space quality was evaluated based on a set of 13 urban quality attributes.Attributes,intended for the street level,were collected from open data platforms and complemented with a field survey.Then,through spatial analysis,these attributes were used to build the Pedestrian Environment Quality Indicator and the Priority Intervention Indicator.Using these indicators,walking-friendly streets were mapped and the most critical sites that require priority intervention were identified to prepare for the requalification planning process.The street with the highest quality in the study area according to the dimensions evaluated-connectivity,convenience,comfort,cleanliness,and conviviality-has a great offer of trees,bus stops,commercial establishments,commercial diversity,crossings,paper bins/eco points and esplanades.On the opposite side of the rank,the street with the lowest quality has no trees or green areas,transport stops,benches and tables,or commercial activity.The urban quality indicators constitute a helpful tool for city planners and policymakers when planning sustainable living neighbourhoods.

Parametric approach to track following control of FFSM

A robust task space tracking scheme is proposed for the free-flying space manipulator system. The dynamic equations of the system are derived via the law of momentum conservation, and then a linear state space representation is formulated by local linearization. A parametric approach is applied by using the eigenstructure assignment theory and the model reference method. A feedback stabilizing controller and a feedforward compensation controller are built based on the approach. Then an optimization procedure is followed after that to obtain the desired requirement and characteristics. Simulation results are presented to show the effectiveness of the proposed method.

Probe of topological invariants using quantum walks of a trapped ion in coherent state space

We present a protocol to realize topological discrete-time quantum walks,which comprise a sequence of spindependent flipping displacement operations and quantum coin tossing operations,with a single trapped ion.It is demonstrated that the information of bulk topological invariants can be extracted by measuring the average projective phonon number when the walk takes place in coherent state space.Interestingly,the specific chiral symmetry owned by our discrete-time quantum walks simplifies the measuring process.Furthermore,we prove the robustness of such bulk topological invariants by introducing dynamical disorder and decoherence.Our work provides a simple method to measure bulk topological features in discrete-time quantum walks,which can be experimentally realized in the system of single trapped ions.

Light pen whole space coordinate measuring system based on a tracking turntable

Light pen coordinate measuring system(LPCMS)is a kind of portable coordinate measuring technique based on vision metrology.In classical LPCMS,the measuring range is limited to the camera’s field of view.To overcome this defect,a new LPCMS is designed in this paper to fulfil whole space coordinate measurement.The camera is installed on a turntable instead of a tripod,so that the camera can rotate to track the movement of the light pen.The new system can be applied to large scale onsite measurement,and therefore it notably extends the application of LPCMS.To guarantee the accuracy of the new system,a method to calibrate the parameters of the tracking turntable is also proposed.Fixing the light pen at a stationary position,and changing the azimuth angles of the turntable’s two shafts,so that the camera can capture the images of the light pen from different view angles.According to the invariant spatial relationship between the camera and the pedestal of the tracking turntable,a system of nonlinear equations can be established to solve the parameters of the turntable.Experimental results show that the whole space coordinate measuring accuracy of the new system can reach 0.25 mm within 10 m.It can be concluded that the newly designed system can significantly expand the measuring range of LPCMS without losing too much accuracy.

Adaptive Sampling for Near Space Hypersonic Gliding Target Tracking

For modern phased array radar systems,the adaptive control of the target revisiting time is important for efficient radar resource allocation,especially in maneuvering target tracking applications.This paper presents a novel interactive multiple model(IMM)algorithm optimized for tracking maneuvering near space hypersonic gliding vehicles(NSHGV)with a fast adaptive sam-pling control logic.The algorithm utilizes the model probabilities to dynamically adjust the revisit time corresponding to NSHGV maneuvers,thus achieving a balance between tracking accuracy and resource consumption.Simulation results on typical NSHGV targets show that the proposed algo-rithm improves tracking accuracy and resource allocation efficiency compared to other conventional multiple model algorithms.

Acquisition,Pointing and Tracking System for Shipborne Space Laser Communication without Prior Information

A space laser communication acquisition,pointing and tracking(APT)system based on the beacon laser is designed without prior information.And then,a new target scanning method and a pointing and tracking algorithm are proposed.The target scanning mode is the round-trip triangular wave scanning,and it means that scanning track of the PAN-TILT platform follows the triangular wave repeatedly.For the pointing and tracking algorithm,the beacon laser is used as the auxiliary aiming light source.The position of the beacon laser in the viewfield of the complementary metal oxide semiconductor(CMOS)camera is calculated by the centroid algorithm.In order to realize the target tracking,the joint control method of the angle control and the angular velocity control is used.The simulation and experimental results show that the APT system can achieve full coverage scanning in the scanning area and capture the target in one scanning cycle successfully.After capturing the PAN-TILT platform,the pointing and tracking algorithm can track the PAN-TILT platform quickly and accurately,and the tracking accuracy is up to 0.22 mrad.

基于在线地图路径循迹的山地城市绿色开放空间步行可达性衰减

【目的】精准客观地评估山地城市绿色开放空间的步行可达性,揭示实际可达性与理论可达性之间的衰减状态,为山地绿色开放空间服务能力优化提升提供依据与参考。【方法】以重庆市渝中区为研究范围,依托ArcGIS平台和高德地图应用程序接口(application programming interface,API),引入路径循迹法,从实际步行路径出发量化可达性,并通过对比实际可达性与理论可达性来分析城市整体及绿色开放空间的可达性衰减状态。【结果】1)渝中区绿色开放空间步行可达范围整体衰减面积比为43.19%,且5 min、10 min、15 min阈值内可达范围的衰减面积比分别为32.12%、46.49%、46.30%;2)绿色开放空间衰减程度不均衡,衰减程度大于60%的绿色开放空间有16个,衰减程度最高为87.71%;3)山顶、各级台层和滨江楔形绿色开放空间的高衰减比例相对更高。【结论】进一步从选址、环境和本体3个方面总结衰减原因,并从步行可达性视角提出增量、串联、渗透的山地城市绿色开放空间优化建议,为客观评估山地城市绿色开放空间步行可达性提供思路,对山地城市绿色开放空间服务效能提升具有重要的现实意义。

虚拟编组下基于跟驰模型的列车群运行控制方法研究

高速铁路移动闭塞方式下运输能力与运输需求之间依然存在缺口,虚拟编组是未来提高运输能力的新方式。本文基于虚拟编组概念,受道路交通跟驰模型启发,引入列车动力学和运动学理论,根据相邻两列车之间的速度和距离关系,提出一种新的追踪列车加速度调整策略,建立相应的列车群虚拟编组加速度控制模型,当实现列车群的虚拟编组时,列车群中所有列车的速度相等,任意相邻列车之间的间距等于期望间距,同时,能够确保列车安全与乘客舒适度。本文以CRH380A动车组为背景验证模型,算例仿真结果表明,本文提出的加速度调整策略能有效实现列车群的虚拟编组;与移动闭塞方式相比,采取虚拟编组方式达到列车协同的时间减少9.7%,列车间距缩小10.1%,效率更高;将列车群视为一个整体时,实现虚拟编组的时间比将整个列车群分为多组时更短。

天空地一体化多目标跟踪算法研究综述

为实现全时全域“泛在连接”,构建天空地一体化网络已成为国家重大需求,而基于天空地一体化网络下跨域协同系统进行多目标跟踪是其中一个重要的发展方向,其在军民用领域都极具应用价值。本文详细阐述了天空地一体化网络背景下多目标跟踪方法研究进展。首先,介绍了天空地一体化跨域协同多目标跟踪的研究背景与意义。其次,从基于视觉的多目标跟踪、基于模型的多目标跟踪和基于多模态融合的多目标跟踪三个方面概述了当前的代表性研究方法:在基于视觉的多目标跟踪算法方面,介绍了单摄像头和多摄像头融合的多目标跟踪方法;对于基于模型的多目标跟踪,先介绍了单传感器多目标跟踪方法,以及在多种复杂场景下的改进,然后介绍了多传感器融合方法;在基于多模态信息融合的目标跟踪方面,在对多传感器时空配准方法和有代表性的多模态信息融合方法介绍的基础上,概括了基于多模态融合的多目标跟踪算法。最后探讨了当前存在的问题和未来发展方向:无论基于视觉的还是基于模型的多目标跟踪方法都有不少问题有待解决,特别是两种方法的结合值得深入研究;在面临复杂干扰时,基于多传感器信息融合的多目标跟踪由于能实现信息的互补,成为未来的主流发展方向;此外,跨域协同系统,由于能利用更多的资源和信息,其多目标跟踪问题研究极具价值,不过其中通信安全问题和多目标跟踪模型轻量化问题值得探讨。本文对从事目标跟踪及空天地一体化协同控制相关理论与技术研究的科研工作者具有重要参考价值。

日常的温暖:“为人”的公共空间

后疫情时代背景下,我国城市建设中尤其需要重视“以人为本”的规划设计,实现环境品质的精细化提升。文章针对快速建设中城市公共空间的规划设计重宏大叙事、轻日常使用的现象,通过问题和症结解析,强调公共空间营造需顺应时代发展下的新需求新动向,突出“为人”的初心,成就“日常温暖”的公共场所,并提出了“径、空、用、场”四项关键原则,即“步行力可及、空间好体验、日常多功用、场所有归属”。

线性自抗扰控制系统的稳定性分析方法

针对二阶线性系统和二阶线性自抗扰控制,提出了基于误差的稳定性分析方法。该方法通过观测误差、跟踪误差状态量建立闭环系统状态空间模型,借助状态空间模型的特征值分析闭环系统的收敛性,并运用李雅普诺夫法判断系统的稳定性。与基于劳斯判据的稳定性分析方法相比较,二者的分析结论一致,算例和仿真结果验证了所提方法的简单性和有效性。此外,通过仿真和蒙特卡洛实验验证了自抗扰控制相比于PID控制在跟踪性能、抗扰能力和鲁棒性等方面的优势。