a) Developed entire gNB 5G uplink mMIMO/multi-user&multi-beam processing algorithm chain, including all algorithm details from mMIMO beamforming, DMRS channel estimation, noise reduction, measurements, interference cancellation, and MIMO equalizer, etc.  Implementable: finite precision, +/-/* only, no Matlab functions Cost-effective: vector computation, efficient for DSP implementation Deployable: balanced & robust in various fading, interference, and packet size. Prototyped in Lab and link performance verified b) Developed entire link simulator for 5G gNB uplink mMIMO/multi-user&multi-beam processing algorithm chain with all the algorithm details, co-developed Lap prototyping Used as O-RAN DU & RU emulator, including mMIMO beamformer Used as UE signal generator, multi-user & multi-layer and fading channel emulator Used as O-RAN DU&RU development platform, test vector generator, debugging, etc.c) Developed a practical solution to self-inflicted PIMC due to RF component non-linearity Practical & competitive: verified competitive in the market on real PIM measured in practical HW setting in lab, and with finite precision & complexity reduction d) 3GPP RAN1 standards, Meta delegator, investigating & Training R18 WI&SI (Advanced 5G)e) Developed 5G PHY training series Physical channels & signals, physical procedures, initial access, antenna array, digital & hybrid beam-forming/tracking, resources allocation, etc. 5G NR baseband signal processing algorithms, beamforming, channel estimation, measurements, noise/interference suppression, multi-user/ multi-layer MIMO equalizer O-RAN split, internal structure of O-DU and O-RUf) Filed 6 patents, and a few more in pipeline for the practical PIMC solution

1. 5G gNB baseband architecture planning on Xilinx RFSoC based platform2. O-RAN fronthaul specification analysis and 5G gNB product planning

a) Proposed and developed (led a team of 50) industry first full-floating point DSP Core & Vector Accelerator for Huawei’s in-house eNB&gNB Baseband SoC, serving billions of people worldwide, with significant increase in performance and power&area efficiency and significant reduction in TTM  Developed a set of novel methodology & tools, derived full-floating point DSP core architecture features & specifications for MIMO&mMIMO baseband signal processing, evaluated the area&power efficiency gain. The key: intimate understanding of the entire baseband signal processing algorithm chain and data flow Led a team of 50 (multi-site/national/functional, over 3 years) that developed & delivered the first in-house DSP subsystem to the eNB&gNB baseband SoC b) Led design teams to develop 4G&5G L1 computation solution & system requirements for the BB SoC Developed a set of methodology for algorithm & SW development that maximizes the potential of the full-floating point DSP based computation system, resulting significant reduction in TTM Proposed/developed user environment (mobility/interference/distance) aware PHY algorithm system that allows intelligent resource allocation, resulting in dramatic capacity increase for BB SoC and choices for trade off. Proposed/developed joint system & algorithm optimization for parallel computation, dramatically increasing HW utilization, reducing internal data traffic/control, and increasing system scalability & SW system reliability c) Invited to work with IF/RF/Ant teams to address frontend impairments, developed a revolutionary solution to PIMC in massive MIMO antenna environments Proposed & developed a baseband solution to PIMC, an key enabler for low-cost & light-weight AAU for mMIMO, requiring a tiny fraction of the resource, as compared with conventional solutiond) Presidential awards: one at company level, another at wireless R&D; and N technological breakthrough awards at wireless R&D; N issued US patents;

a) (95-99) Developed & delivered two-way paging basestation baseband receiver: signal processing algorithms and SW architecture (including real-time operating system). 5dB better than competitors, caused a “big problem” for system deployment, resulting in modification of system deployment planning in a good way; with complete Matlab simulationb) (99-01) Developed and delivered Bluetooth baseband receiver: signal processing algorithms with complete Matlab link simulation, supporting ASIC implementation, 5dB better performance than competitors, a key patent (US7,639,762B2) widely used according to industry experts, especially for LEO satellite communication c) (01-02) Developed WiFi baseband receiver algorithms for ASIC implementation, with complete Matlab based link simulator.d) (03-09) Renovated Motorola’s 2G/2.5G UE baseband signal processing algorithms, implemented 50% SW for 2G/2.5G core signal processing algorithms (micro-vector processor), with complete Matlab link simulatione) (05-09) Freescale delegator to 3GPP GRANf) Received 30 issued US patents, N awards; Distinguished Member of Technical Staff