Neuromorphic interconnects using Ultra Wideband radio

We present the architecture and testing of an address event representation (AER) wireless link that uses ultra wideband (UWB) radio as the physical layer. The wireless protocol layers from high to low level are: AER events, asynchronous handshaking, PRN coding, and UWB transmission. The top three layers are implemented in digital logic in an FPGA, while the UWB physical layer is implemented using discrete RF components. The combination of error detection during asynchronous handshaking and PRN codes add noise robustness to the wireless link. First we describe the implementation and testing of the UWB AER link, demonstrating the operation of the protocol. Second we analyze the capacity of a network of UWB AER nodes, finding it able to support several thousand neurons across multiple transmitters, even at relatively high sustained firing rates.

[1]  William Bialek,et al.  Spikes: Exploring the Neural Code , 1996 .

[2]  Misha Anne Mahowald,et al.  VLSI analogs of neuronal visual processing: a synthesis of form and function , 1992 .

[3]  Eugenio Culurciello,et al.  CMOS image sensors for sensor networks , 2006 .

[4]  Bertram E. Shi,et al.  Neuromorphic implementation of orientation hypercolumns , 2005, IEEE Transactions on Circuits and Systems I: Regular Papers.

[5]  Andreas G. Andreou,et al.  Address-data event representation for communication in multichip neuromorphic system architectures , 2007 .

[6]  E. Culurciello,et al.  A biomorphic digital image sensor , 2003, IEEE J. Solid State Circuits.

[7]  Kwabena Boahen.  Retinomorphic vision systems , 1996, Proceedings of Fifth International Conference on Microelectronics for Neural Networks.

[8]  Kwabena Boahen,et al.  Point-to-point connectivity between neuromorphic chips using address events , 2000 .

[9]  A. Cassidy,et al.  FPGA Based Silicon Spiking Neural Array , 2007, 2007 IEEE Biomedical Circuits and Systems Conference.

[10]  André van Schaik,et al.  AER EAR: A Matched Silicon Cochlea Pair With Address Event Representation Interface , 2005, IEEE Transactions on Circuits and Systems I: Regular Papers.

[11]  Moe Z. Win,et al.  Impulse radio: how it works , 1998, IEEE Communications Letters.

[12]  Gert Cauwenberghs,et al.  Probabilistic synaptic weighting in a reconfigurable network of VLSI integrate-and-fire neurons , 2001, Neural Networks.

[13]  Giacomo Indiveri,et al.  A VLSI reconfigurable network of integrate-and-fire neurons with spike-based learning synapses , 2004, ESANN.

[14]  Gert Cauwenberghs,et al.  A Multichip Neuromorphic System for Spike-Based Visual Information Processing , 2007, Neural Computation.

[15]  M.B. Pursley,et al.  Crosscorrelation properties of pseudorandom and related sequences , 1980, Proceedings of the IEEE.

[16]  Massimo A. Sivilotti,et al.  Wiring considerations in analog VLSI systems, with application to field-programmable networks , 1992 .