High Resolution binocular MIMO Radars at Millimeter Wave Frequencies (binoMIMO)

未来，毫米波频率的雷达传感器将广泛用于各种应用，如汽车、自动化技术、安全工程、物流、质量控制或生物医学应用。由于硅锗技术的进步，紧凑、节能和经济高效的雷达传感器可以在100 GHz以上的毫米波频率范围内实现。由于工作频率高，系统带宽大，同时可以用紧凑的雷达传感器物理尺寸实现波长倍数的非常大的孔径。然而，这些成像系统在高频信号的分布、高集成度导致的非理想性和高效标定等方面存在着主要问题。在“高分辨率双目MIMO毫米波雷达”项目中，首次将新型MIMO雷达概念与基于模型的信号处理方法相结合。为此目的，乌尔姆大学在毫米波雷达技术领域的专业知识和伊尔梅瑙工业大学在高分辨率参数估计领域的专业知识相结合。建立了一种新型MIMO雷达系统，该系统由两个耦合MIMO雷达组成，工作频率约为160 GHz，可实现近距离分布式目标的高分辨率双目或准立体观测。分析了工作在FMCW原理下的雷达的分辨率和噪声特性。采用基于模型的信号处理方法补偿硬件缺陷，提高了雷达的分辨率和鲁棒性。本文的研究方向是利用高频信号的非相干分布来降低双目雷达的硬件复杂度。在参数估计器中应考虑硬件的这些适应性。在此基础上，建立雷达目标模型（目标模型）并将其集成到估计器中。这为物体的分类提供了更多的可能性，也为确定物体的位置和方向提供了可能性。由于阵列孔径较大，在参数估计中还应考虑球面波前。为了在参数估计器中考虑柔性双目排列的阵列模型，还必须开发有效的校准程序。乌尔姆大学和伊尔梅瑙工业大学之间的合作保证了双目雷达系统的进一步发展及其理论和实验验证。

In the future, radar sensors for millimeter wave frequencies will be used in a wide variety of applications such as automotive, automation technology, safety engineering, logistics, quality control or biomedical applications. Thanks to progress in silicon-germanium technology, compact, energy-efficient and cost-effective radar sensors can be implemented at frequencies in the millimeter wave range above 100 GHz. Due to the high operating frequencies, large system bandwidths and at the same time very large apertures in multiples of the wavelength can be realized with compact physical dimensions of the radar sensor. However, there are major problems with the distribution of the high-frequency signal, the non-idealities due to the high integration level and the efficient calibration of these imaging systems.In the project "High-resolution binocular MIMO millimeter wave radars", a novel MIMO radar concept was combined with model-based signal processing methods for the first time. The expertise of the University of Ulm in the field of millimeter wave radar technology and the expertise of the TU Ilmenau in the field of high-resolution parameter estimation was combined for this purpose. A novel MIMO radar system consisting of two coupled MIMO radars operating at approximately 160 GHz was built, which allows high-resolution binocular or quasi-stereoscopic observation of distributed objects at a close range. The radar operating with the FMCW principle has been analyzed with respect to its resolution and noise properties. Methods of model-based signal processing were used to compensate for hardware impairments, such that the resolution and robustness of the radar has been enhanced.This continuative research proposal deals with the reduction of the hardware complexity of the binocular radar by a non-coherent distribution of the high-frequency signal. These adaptations in the hardware shall be considered in the parameter estimator. Furthermore, a model of the radar target (object model) is to be developed and integrated into the estimator. This offers more possibilities for the classification of objects, as well as the possibility to determine position and orientation of the objects. Due to the large array aperture, spherical wave fronts should also be considered in the parameter estimator. In order to consider an array model for the flexible binocular arrangement within the parameter estimator, efficient calibration procedures have to be developed, too.The cooperation between the University of Ulm and the TU Ilmenau guarantees the further development of the binocular radar system as well as its theoretical and experimental verification.