Mars aerocapture is one of the most concerned technologies for future Mars sample-return and manned exploration missions. This article investigates the challenging problem of the reference trajectory tracking guidance for Mars aerocapture under uncertainties. Based on an integral sliding surface, a fixed-time neural adaptive tracking guidance law is synthesized by incorporating the neural network (NN) approximation into the fixed-time integral sliding mode control approach. Benefiting from the integral sliding surface design, the proposed tracking guidance law has no singularity problem inherently existing in terminal sliding mode control. By adopting the NN approximation to compensate for the lumped uncertain term in the feedforward loop, the proposed tracking guidance law is strongly robust against aerodynamic coefficient uncertainties and atmospheric density uncertainty. Stability analysis shows that the radial distance tracking error and its time derivative can stabilize to the small neighborhoods around the origin in fixed time under the proposed tracking guidance law. Finally, the effectiveness and advantages of the proposed tracking guidance law are illustrated through simulations and comparisons.

中文翻译：

---

具有神经补偿的火星航空捕获非奇异固定时间跟踪制导


火星航空捕获是未来火星采样返回和载人探测任务最受关注的技术之一。本文研究了不确定条件下火星航空捕获参考轨迹跟踪制导的挑战性问题。基于积分滑模面，将神经网络（NN）近似融入固定时间积分滑模控制方法中，合成了固定时间神经自适应跟踪制导律。得益于整体滑模面设计，所提出的跟踪制导律不存在终端滑模控制固有的奇异性问题。通过采用神经网络近似来补偿前馈回路中的集总不确定项，所提出的跟踪制导律对于空气动力系数不确定性和大气密度不确定性具有很强的鲁棒性。稳定性分析表明，在所提出的跟踪制导律下，径向距离跟踪误差及其时间导数能够在固定时间内稳定到原点周围的小邻域。最后，通过仿真和比较说明了所提出的跟踪制导律的有效性和优点。