It has been seen that the satellite industry is changing from being a manufacturing industry to being a service providing industry. The main contributors of the cost of providing satellite services will start to move from manufacturing costs towards software and intellectual-property development costs. In this sense, the manufacturing and launching costs are expected to decrease due to standardization and mass production, while the implementation of small satellites and nano-satellites is already gaining momentum.It has been foreseen that the future high demand for communications and earth observation, that comes with the IoT and MoM era (Internet-of-Things and Machine-to-Machine communications), can be covered by arrays of small satellites (e.g. Orbcomm, ClydeSpace) instead of single big platforms. Some examples are the projects OneWeb, Leosat and SpaceX, which plan micro satellite chain production, in order to deploy mega constellations that expect to include thousands of satellites.Another slightly different approach to the mega constellations, are the satellite swarms in close formation. In this topology, a ground-based target can be pointed from signals generated concurrently from different platforms. This function performed by the array is known as distributed beamforming (or distributed synthetic aperture array operation). These technologies, despite they have being extensively studied by academic and industrial institutions, have not been taken into the practice to real satellite implementations due to the difficulty that implies the required coordination of the multiple satellites. This coordination includes synchronization in time, frequency, phase, and the accurate localization of the satellites in the space over time, and their angular attitudes. The synchronization of digital radios can be seen as a mature field; however, there is not currently a framework that includes all the absolute synchronization and localization requirements that allow the aforementioned implementations. In addition to the lack of accurate synchronization, the exact orbits followed by satellites are only known a-posteriori, where rapid response times are required by the distributed beam-forming algorithms.In this project, we propose the use of a multi-band bidirectional inter-satellite coordination links. These links are designed using microwave/RF techniques, advanced data processing, statistical estimation and active learning. The proposed coordination schemes will be studied analytically and validated by means of software defined radio techniques on physical radio platforms in laboratory and in outdoor experiments.The proposed architecture will provide dynamically the accurate navigation (inter-satellite distance) and synchronization required to the implementation of cohesive constellations. Whereas, this cohesiveness implies that the whole constellation synthesizes as a single entity with an antenna of immense dimensions. This will open the possibility to exiting new applications that have not even been thought before, like high-resolution microwave radiometry, for the observation of the earth and above the earth for the detection of asteroids or other astronomy applications. As well, in the field of communications, it can be applied in the interference localization and mitigation techniques, and in the implementation of ubiquitous low latency internet connections with a global coverage, in which the hand-over procedures are replaced by a gradual change in the weighting of a distributed beamforming process.