List of innovations
1. Hybrid Network Manager components and functionalities as a backhauling improvement proposal
This innovation proposes a novel network management scheme for hybrid terrestrial/satellite backhaul networks. This scheme aims for a self-organizing and flexible solution capable of efficiently routing the mobile traffic in terms of capacity and energy consumption, while providing resilience against link failures or congestion and easy deployment in rural areas. In order to cover these requirements, two key elements have been proposed: a centralized Hybrid Network Manager (HNM) and distributed intelligent backhaul nodes (iBNs). The iBN extends the internal architecture of traditional BNs by introducing new functional blocks and interfaces for the proper management of backhaul satellite and terrestrial resources, such as hybrid routing, traffic classification or an energy efficiency agent. In other hand, the HNM, based on a global network view, performs topology calculations for adapting to traffic demands as well as QoS and energy efficiency management.
The HNM-iBN pair has been conceived with many important functionalities that could be separately exploited. The three innovations to be exploited as part of the HNM/iBN are listed below:
1a. Distributed routing algorithm for hybrid terrestrial-satellite backhaul networks
This innovation proposes a distributed routing algorithm which is able to integrate seamlessly the satellite and terrestrial network resources while providing adaptation in terms of topology changes and network load conditions. This algorithm is based on the backpressure concept and it is able to take forwarding decisions with different granularities. The benefits of the proposed backpressure algorithm has been demonstrated by simulations under the different scenarios defined in the SANSA project and have been already published. In terms of capacity, the routing algorithm showed up to 60% of improvements compared to SoA whereas latency is reduced by up to 3x with our backpressure algorithm.
1b. Distributed Q-learning for energy harvesting heterogeneous networks
This innovation proposes small cells powered with solar panels and controlled with decentralized SON solution based on multi-agent reinforcement learning (RL). The proposed solution enables a potential energy savings of up to 50% respect to the case in which small cells are powered with the grid. Moreover, the Q-learning algorithm proposed provides up to 16% improvements both in terms of throughput and energy efficiency with respect to standard baseline solutions. This is obtained thanks to the learning behaviour, which allows to adapt the switch ON/OFF policies according to energy harvested income and the traffic profiles. Finally, the algorithm presented provides a harvested energy surplus, which can be used by mobile and/or infrastructure operators to offer ancillary services to the smart electricity grid. The main properties and the benefits of the proposed algorithm has been demonstrated by simulations and already published.
1c. Link scheduling and flow assignment for hybrid satellite-terrestrial backhaul networks
This innovation proposes a link scheduling method to tackle radio and interference limitations in self-organizing backhaul networks. The solution enables aggressive frequency reuse and an efficient use of the radio equipment. Extension to the case of having several frequency bands is also considered. The work is already published.
2. Hybrid A/D beamforming with interference constraints
This innovation consist of a set of algorithms for determining the beamforming matrices in hybrid analog-digital arrays which permit establishing the desired directive links while limiting the interference to other sensible directions. This feature is of vital importance for future 5G mm-wave systems alone as well as for the coexistence of terrestrial and satellite links on the same bands. The innovation could be of interest for antenna array manufacturers as Viasat as well as for chip manufacturers in the beamforming domain.
This innovation groups the five different innovative beamforming algorithms detailed below. The joint exploitation of those algorithms will permit addressing different customer requirements.
2a. Hybrid A/D beamforming for Cognitive mm-wave Systems
This innovation proposes low complexity algorithms that derive the optimal beamforming matrices that enable coexistence for large array systems in mm-wave bands. The algorithms are already derived and application in Terrestrial-to-Terrestrial links is already done.
2b. Hybrid analog-digital beamforming strategies for massive-MIMO mm-wave backhaul using sub-connected antenna arrays
This innovation proposes a sub-connected hybrid digital-analog BF technique for wireless backhaul networks which exploits the best beam approach. A two-stage Zero Forcing approach is employed to steer beams and nulls into predefined directions according to the known topology, which renders CSI estimation unnecessary. Parameters of the proposed BF architecture, e.g. resolution of phase shifters or allocation of antenna elements to RF chains, are optimized with respect to array gain and the interference suppression, which are the most important metrics for high quality communication backhaul links. The algorithm was demonstrated by simulations and will be(was) submitted for publication to “IEEE JSTSP Special Issue on Hybrid Analog - Digital Signal Processing for Hardware-Efficient Large Scale Antenna Arrays”.
2c. Phase-only transmit and receive beamforming design with interference control
The optimization method for dealing with phased array beamforming conceived in SANSA is novel. The proposed scheme is the only one able to perfectly control the total interference power towards a set of directions.
2d. Hybrid analog-digital beamforming with interference control
The proposed alternating optimization method differs to the current approaches and it is able to control the interference towards non-intended receivers. The mechanism is able to tackle beamforming networks consisting of switches, phase-shifters assuming an arbitrary connectivity matrix.
2e. Hybrid analog-digital multiuser precoding with QoS guarantees
Controlling the SINR values of different users with an underlaying hybrid analog-digital scheme is proposed in here. This approach contrasts with current works on fully-digital solutions which are unable to design the analog part.
3. Peak power minimization in symbol-level precoding for cognitive MISO downlink channels
This innovation proposes a solution for mitigating the non-linearities of the RF amplifiers in the downlink of cognitive multiuser MISO system, based on reducing the antennas power variance to limit the differential phase distortion through symbol-level precoding. The algorithms have been demonstrated by simulations in the terrestrial and satellite scenarios. A patent is already covering part of the innovation. In addition, a First Lab demo has been achieved and is demonstrated here.
4. Dual beam Ka band antenna array for Satellite-Terrestrial backhaul networks
This innovation consist of a low profile Ka band antenna array for Satellite-Terrestrial backhaul networks. This new solution allows to dynamically point the beam, to mitigate interference and increases the capacity and flexibility of the backhaul network. The antenna is manufactured using the process described in the provisional patent (#60102) allowing for the suppression of surface wave mode with a cost efficient technique.
5. Parasitic antenna arrays
5a. 19.25GHz multi-active multi-passive antenna array
This innovation proposes a new planar antenna design, operating at the 19.25GHz frequency band regime that has 4 active and 40 parasitic elements. It presents good simulation results in terms of beam shaping and beam steering, by varying the values of the capacitors which are connected to the parasitic elements. Although the overall array gain is not as high as in other hybrid arrays, its low-cost and easy-to-build characteristics make it a plausible future candidate for terrestrial communication.
5b. Hybrid SPSA stochastic beam forming algorithm
This innovation presents a new algorithm, the hybrid stochastic-deterministic algorithm, for the computation of the loads for the passive elements and the baseband weights. The benefit of this algorithm is that it produces more accurate load values and baseband weights values, leading to better estimation of the desired radiation pattern. This algorithm enables beam shaping and beam steering in the presence of mutual coupling, whereas other existing works neglect the mutual coupling effects.
6. Caching algorithms
6a. Novel hybrid reactive caching algorithm
This innovation proposes a novel reactive caching algorithm called Hybrid Score-Gated Least Recently Used (SG-LRU). This algorithm maintains the benefits of LRU (simple implementation, low computational burden, and adaptation to content popularity dynamics), while at the same time it improves significantly the achieved cache hit rate and undercuts the loading rate of content into the cache, in comparison with standard LRU.
6b. Novel cooperative reactive caching algorithms
This innovation presents a family of cooperative reactive caching schemes based on the SG-LRU algorithm. These variants outperform the cooperative LRU strategy and reduce the cooperation overhead.
7. Row-column beam-steering control in reflectarrays
This innovation proposes a new reflectarray cell design that enables a control scheme based on rows and columns which results in a significant complexity reduction with respect to traditional per-element control. The new design performance has been demonstrated by simulations and already published.
8. Linear baseband precoding strategies for millimeter wave MIMO multi-X channel
This innovation proposes low-complexity linear baseband precoding and combining schemes for a MIMO multi-X channel based on large antenna arrays. These new precoding strategies mitigate the multi-user interference which increases the sum-rate (throughput) of the communications system. The algorithms were demonstrated by simulations and already published.
9. Design method for arbitrary multi-beam leaky-wave antennas
This method combines holographic and superposition principles for the synthesis of arbitrary multi-beam LWA, including the synthesis of more than two beams or different beams with different gains. It is included in a paper accepted for publication in PIERS 2017 Singapore.
10. Design method for nulling in multi-feed metasurface antennas
This method consist of synthesizing contiguous beams with different beams and combine them digitally to suppress undesired interferences. It has been simulated, demonstrated experimentally and included in D5.1.
11. Content distribution optimization for hybrid Satellite-Terrestrial backhaul networks
This innovation proposes a new algorithm to be included in the SANSA Hybrid Network Manager in order to determine the cache contents more prioritize, to predict the traffic requirements and in turn permitting optimize the network traffic around at least in 30%. The algorithm will be based on analysing the social networks tendencies and information from main application distribution portals. This algorithm has been conceived but not developed yet.
12. Traffic policies for hybrid Satellite-Terrestrial backhaul networks under emergency situations
This innovation proposes a method for the optimization and real time control over traffic performance in emergency situations with autonomous operation. It would be in charge of, disabling traffic sources not related to voice interchange application, prioritizing messages and voice data, distributing the traffic using the communications available in function of the flows and capacities, etc. The method has been conceived to relay on the SANSA architecture based on the Hybrid Network Manager and intelligent Backhaul Nodes, but still needs to be developed.