Modeling irregular small bodies gravity field via extreme learning machines and Bayesian optimization

Close proximity operations around small bodies are extremely challenging due to their uncertain dynamical environment. Autonomous guidance and navigation around small bodies require fast and accurate modeling of the gravitational field for potential on-board computation. In this paper, we investigate a model-based, data-driven approach to compute and predict the gravitational acceleration around irregular small bodies. More specifically, we employ Extreme Learning Machine (ELM) theories to design, train and validate Single-Layer Feedforward Networks (SLFN) capable of learning the relationship between the spacecraft position and the gravitational acceleration. ELM-base neural networks are trained without iterative tuning therefore dramatically reducing the training time. Analysis of performance in constant density models for asteroid 25143 Itokawa and comet 67/P Churyumov-Gerasimenko show that ELM-based SLFN are able learn the desired functional relationship both globally and in selected localized areas near the surface. The latter results in a robust neural algorithm for on-board, real-time calculation of the gravity field needed for guidance and control in close-proximity operations near the asteroid surface.     

Modeling and experimental validation of sawing based lander anchoring and sampling methods for asteroid exploration

This paper presents a novel lander anchoring system based on sawing method for asteroid exploration. The system is composed of three robotic arms, three cutting discs, and a control system. The discs mounted at the end of the arms are able to penetrate into the rock surface of asteroids. After the discs cut into the rock surface, the self-locking function of the arms provides forces to fix the lander on the surface. Modeling, trajectory planning, simulations, mechanism design, and prototype fabrication of the anchoring system are discussed, respectively. The performances of the system are tested on different kinds of rocks, at different sawing angles, locations, and speeds. Results show that the system can cut 15?mm deep into granite rock in 180?s at sawing angle of 60°, with the average power of 58.41?W, and the “weight on bit” (WOB) of 8.637?N. The 7.8?kg anchoring system is capable of providing omni-directional anchoring forces, at least 225?N normal and 157?N tangent to the surface of the rock. The system has the advantages of low-weight, low energy consumption and balance forces, high anchoring efficiency and reliability, and could enable the lander to move and sample or assist astronauts and robots in walking and sampling on asteroids.     

双星系统L1点悬停探测控制器设计与仿真

双小行星系统探测具有重要的科学意义,受其复杂动力学环境影响,探测任务极具挑战。利用球谐函数法对双星系统进行引力场建模,求解双星系统平动点,并选取其内部共线平动点L_1点作为双星系统悬停探测目标位置。采用航天探测实际任务中常用的脉冲推力式发动机,设计了一种原理简单、便于工程实现的常值切换bang-bang控制器。以69230Hermes双星系统为例,将Hermes近似为双椭球系统,仿真分析航天器在Hermes双星系统L_1点悬停飞行的控制效果,验证所提控制策略有效性。     

Polyhedron tracking and gravity tractor asteroid deflection

In the wake of the Chelyabinsk airburst, the defense against hazardous asteroids is becoming a topic of high interest. This work improves the gravity tractor asteroid deflection approach by tracking realistic small body shapes with tilted ion engines. An algorithm for polyhedron tracking was evaluated in a fictitious impact scenario. The simulations suggest a capability increase up to 38.2% with such improved tilting strategies. The long- and short-term effects within polyhedron tracking are illustrated. In particular, the orbital reorientation effect is influential when realistic asteroid shapes and rotations are accounted for. Also analyzed is the subject of altitude profiles, a way to tailor the gravity tractor performance, and to achieve a steering ability within the B-plane. A novel analytical solution for the classic gravity tractor is derived. It removes the simulation need for classic tractor designs to obtain comparable two body model Δv$\mathrm{\Delta }v$ figures. This paper corroborates that the asteroid shape can be exploited for maximum performance. Even a single engine tilt adjustment at the beginning of deflection operations yields more deflection than a fixed preset tilt.     

Rapid design and optimization of low-thrust rendezvous/interception trajectory for asteroid deflection missions

Asteroid deflection techniques are essential in order to protect the Earth from catastrophic impacts by hazardous asteroids. Rapid design and optimization of low-thrust rendezvous/interception trajectories is considered as one of the key technologies to successfully deflect potentially hazardous asteroids. In this paper, we address a general framework for the rapid design and optimization of low-thrust rendezvous/interception trajectories for future asteroid deflection missions. The design and optimization process includes three closely associated steps. Firstly, shape-based approaches and genetic algorithm (GA) are adopted to perform preliminary design, which provides a reasonable initial guess for subsequent accurate optimization. Secondly, Radau pseudospectral method is utilized to transcribe the low-thrust trajectory optimization problem into a discrete nonlinear programming (NLP) problem. Finally, sequential quadratic programming (SQP) is used to efficiently solve the nonlinear programming problem and obtain the optimal low-thrust rendezvous/interception trajectories. The rapid design and optimization algorithms developed in this paper are validated by three simulation cases with different performance indexes and boundary constraints.     

Dynamics and control of spacecraft hovering using the geomagnetic Lorentz force

To achieve hovering, a spacecraft thrusts continuously to induce an equilibrium state at a desired position. Due to the constraints on the quantity of propellant onboard, long-time hovering around low-Earth orbits (LEO) is hardly achievable using traditional chemical propulsion. The Lorentz force, acting on an electrostatically charged spacecraft as it moves through a planetary magnetic field, provides a new propellantless method for orbital maneuvers. This paper investigates the feasibility of using the induced Lorentz force as an auxiliary means of propulsion for spacecraft hovering. Assuming that the Earth’s magnetic field is a dipole that rotates with the Earth, a dynamical model that characterizes the relative motion of Lorentz spacecraft is derived to analyze the required open-loop control acceleration for hovering. Based on this dynamical model, we first present the hovering configurations that could achieve propellantless hovering and the corresponding required specific charge of a Lorentz spacecraft. For other configurations, optimal open-loop control laws that minimize the control energy consumption are designed. Likewise, the optimal trajectories of required specific charge and control acceleration are both presented. The effect of orbital inclination on the expenditure of control energy is also analyzed. Further, we also develop a closed-loop control approach for propellantless hovering. Numerical results prove the validity of proposed control methods for hovering and show that hovering around low-Earth orbits would be achievable if the required specific charge of a Lorentz spacecraft becomes feasible in the future. Typically, hovering radially several kilometers above a target in LEO requires specific charges on the order of 0.1 C/kg.     

Methods of spacecraft impulsive relative hovering and trajectory safety analysis

Based on the analytical solutions of T-H equations and its state transition matrix form,the open-loop control method of spacecraft impulsive relative hovering was studied,which is promising for practical engineering use.The true anomaly intervals of the hovering impulse were optimized by the nonlinear mathematical programming.Based on the calculation of collision probability,the method of safety analysis and risk management was proposed.The numerical simulations show that the introduced relative hovering method can be used for circular and elliptical reference orbits hovering.Furthermore,the local optimal solution can be obtained by applying the true anomaly intervals optimization method.The maximum collision probability and the minimum relative distance nearly appear at the same time.And,the smaller the relative distance is,the larger the collision probability.     

用参数辨识法进行直升机悬停性能拓展

准确计算直升机在不同飞行状态的有效气动参数对于确定直升机飞行性能具有重要意义。然而,由于复杂的旋翼空气动力现象以及直升机状态和环境条件的变化,准确预估气动参数有较大难度。为此,采用无量纲分析法建立直升机悬停状态的数学模型,首先对参数重组,确定了几个悬停状态重要参数,包括气动参数和直升机状态参数;然后,以直-9×型直升机为例,结合实际试飞数据,提出了用最小二乘法对该模型进行参数辨识的方法;最后,通过相关性分析,确定了辨识方法的可行性,并将辨识结果有效地用于直升机悬停性能拓展。结果表明,这种利用参数辨识进行性能拓展的方法是可行的,由于辨识结果是利用实际试飞数据确定的,拓展结果具有较高的可信度。这种数据处理方法可有效减少试飞周期,节约试验成本。     

电磁航天器编队悬停鲁棒协同控制方法

针对电磁航天器编队近地轨道悬停问题,提出一种在缺少参考轨道准确信息时的协同控制方法。用TH方程描述航天器间的相对运动,选择与参考轨道同周期的圆轨道为标称轨道。将参考轨道相对于标称圆轨道的偏差、地球非球形引力、大气阻力及其他天体引力等参数单独归类,视其为不确定量,构成不确定系统。通过引入一致性理论,在电磁作用模型和动力学方程均存在不确定性的条件下,针对航天器编队悬停的目标设计了鲁棒协同控制律。考虑能量消耗最优和均衡以及轨道姿态解耦,给出了通过优化进行磁矩配置的方案。仿真结果表明,所设计的鲁棒协同控制律能够实现编队电磁航天器高精度悬停,所给出的磁矩配置方案能够实现磁矩的合理分配。      

弱引力下小行星动量驱动机器人的可控跳跃行为

以小行星表面着陆探测为背景,提出一种动量驱动机器人(MoRo)以满足弱引力复杂环境下的探测需求。该机器人利用弱引力环境下的摩擦和碰撞特性,通过主动辨识环境参数,规划和控制动量轮以产生期望的驱动力矩,完成可控性跳跃及腾空后的稳定拍照等任务。首先,基于MoRo的动量轮刹车机构特性,分析了MoRo在弱引力环境下的跳跃机理并对其跳跃方式进行了规划;接着考虑动量轮驱动机构三闭环伺服系统的非线性特性,基于Herze碰撞模型和Karnopp摩擦模型建立了MoRo在小行星表面的跳跃行为动力学模型;其次,使用机器学习算法建立环境参数和MoRo运动的函数关系,并基于环境参数规划动量轮转速实现跳跃距离和腾空高度的可控。最后,通过数值仿真校验了MoRo跳跃规划方法和控制方法的可行性。