{"id":"https://openalex.org/W4416748996","doi":"https://doi.org/10.1109/iros60139.2025.11247019","title":"Computationally Efficient FPGA-based Large Language Model Inference for Real-Time Decision-Making in Robotic Systems","display_name":"Computationally Efficient FPGA-based Large Language Model Inference for Real-Time Decision-Making in Robotic Systems","publication_year":2025,"publication_date":"2025-10-19","ids":{"openalex":"https://openalex.org/W4416748996","doi":"https://doi.org/10.1109/iros60139.2025.11247019"},"language":null,"primary_location":{"id":"doi:10.1109/iros60139.2025.11247019","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iros60139.2025.11247019","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101646695","display_name":"Huaizhi Zhang","orcid":"https://orcid.org/0000-0003-0338-1497"},"institutions":[{"id":"https://openalex.org/I110002522","display_name":"University of Essex","ror":"https://ror.org/02nkf1q06","country_code":"GB","type":"education","lineage":["https://openalex.org/I110002522"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Huaizhi Zhang","raw_affiliation_strings":["University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ","institution_ids":["https://openalex.org/I110002522"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072305339","display_name":"Tamim M. Al-Hasan","orcid":"https://orcid.org/0000-0002-9927-4305"},"institutions":[{"id":"https://openalex.org/I110002522","display_name":"University of Essex","ror":"https://ror.org/02nkf1q06","country_code":"GB","type":"education","lineage":["https://openalex.org/I110002522"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Tamim M. Al-Hasan","raw_affiliation_strings":["University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ","institution_ids":["https://openalex.org/I110002522"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101835782","display_name":"Xuqi Zhu","orcid":"https://orcid.org/0009-0007-1131-6237"},"institutions":[{"id":"https://openalex.org/I110002522","display_name":"University of Essex","ror":"https://ror.org/02nkf1q06","country_code":"GB","type":"education","lineage":["https://openalex.org/I110002522"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Xuqi Zhu","raw_affiliation_strings":["University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ","institution_ids":["https://openalex.org/I110002522"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072271555","display_name":"Jiacheng Zhu","orcid":"https://orcid.org/0000-0001-6990-5228"},"institutions":[{"id":"https://openalex.org/I110002522","display_name":"University of Essex","ror":"https://ror.org/02nkf1q06","country_code":"GB","type":"education","lineage":["https://openalex.org/I110002522"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Jiacheng Zhu","raw_affiliation_strings":["University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ","institution_ids":["https://openalex.org/I110002522"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002861905","display_name":"Weiyong Si","orcid":"https://orcid.org/0000-0003-4531-2596"},"institutions":[{"id":"https://openalex.org/I110002522","display_name":"University of Essex","ror":"https://ror.org/02nkf1q06","country_code":"GB","type":"education","lineage":["https://openalex.org/I110002522"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Weiyong Si","raw_affiliation_strings":["University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ","institution_ids":["https://openalex.org/I110002522"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077992830","display_name":"Klaus D. McDonald-Maier","orcid":"https://orcid.org/0000-0002-6412-8519"},"institutions":[{"id":"https://openalex.org/I110002522","display_name":"University of Essex","ror":"https://ror.org/02nkf1q06","country_code":"GB","type":"education","lineage":["https://openalex.org/I110002522"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Klaus D. McDonald-Maier","raw_affiliation_strings":["University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ","institution_ids":["https://openalex.org/I110002522"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5042584242","display_name":"Xiaojun Zhai","orcid":"https://orcid.org/0000-0002-1030-8311"},"institutions":[{"id":"https://openalex.org/I110002522","display_name":"University of Essex","ror":"https://ror.org/02nkf1q06","country_code":"GB","type":"education","lineage":["https://openalex.org/I110002522"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Xiaojun Zhai","raw_affiliation_strings":["University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Essex,School of Computer Science and Electronic Engineering,Colchester,United Kingdom,CO4 3SQ","institution_ids":["https://openalex.org/I110002522"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I110002522"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.42375796,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"13141","last_page":"13146"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.4124999940395355,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.4124999940395355,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.18970000743865967,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T14347","display_name":"Big Data and Digital Economy","score":0.061500001698732376,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.7531999945640564},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5776000022888184},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.5164999961853027},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.4864000082015991},{"id":"https://openalex.org/keywords/implementation","display_name":"Implementation","score":0.4657000005245209},{"id":"https://openalex.org/keywords/computational-model","display_name":"Computational model","score":0.3831000030040741},{"id":"https://openalex.org/keywords/constraint","display_name":"Constraint (computer-aided design)","score":0.35109999775886536},{"id":"https://openalex.org/keywords/efficient-energy-use","display_name":"Efficient energy use","score":0.3395000100135803},{"id":"https://openalex.org/keywords/optimization-problem","display_name":"Optimization problem","score":0.3190000057220459}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7699999809265137},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.7531999945640564},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5776000022888184},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.5164999961853027},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.4864000082015991},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.47110000252723694},{"id":"https://openalex.org/C26713055","wikidata":"https://www.wikidata.org/wiki/Q245962","display_name":"Implementation","level":2,"score":0.4657000005245209},{"id":"https://openalex.org/C66024118","wikidata":"https://www.wikidata.org/wiki/Q1122506","display_name":"Computational model","level":2,"score":0.3831000030040741},{"id":"https://openalex.org/C2776036281","wikidata":"https://www.wikidata.org/wiki/Q48769818","display_name":"Constraint (computer-aided design)","level":2,"score":0.35109999775886536},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.3395000100135803},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.32350000739097595},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3199000060558319},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.3190000057220459},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.31520000100135803},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.3107999861240387},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.29899999499320984},{"id":"https://openalex.org/C42143788","wikidata":"https://www.wikidata.org/wiki/Q173341","display_name":"Hardware description language","level":3,"score":0.29499998688697815},{"id":"https://openalex.org/C90509273","wikidata":"https://www.wikidata.org/wiki/Q11012","display_name":"Robot","level":2,"score":0.2867000102996826},{"id":"https://openalex.org/C49154492","wikidata":"https://www.wikidata.org/wiki/Q5300","display_name":"Central processing unit","level":2,"score":0.27900001406669617},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.27880001068115234},{"id":"https://openalex.org/C26771161","wikidata":"https://www.wikidata.org/wiki/Q16980","display_name":"ARM architecture","level":2,"score":0.2752000093460083},{"id":"https://openalex.org/C68781425","wikidata":"https://www.wikidata.org/wiki/Q2052203","display_name":"Multi-objective optimization","level":2,"score":0.27489998936653137},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.26570001244544983},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.26420000195503235},{"id":"https://openalex.org/C186370098","wikidata":"https://www.wikidata.org/wiki/Q442787","display_name":"Energy (signal processing)","level":2,"score":0.2612999975681305},{"id":"https://openalex.org/C2779439875","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Natural language understanding","level":3,"score":0.260699987411499},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.259799987077713},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.25760000944137573},{"id":"https://openalex.org/C34413123","wikidata":"https://www.wikidata.org/wiki/Q170978","display_name":"Robotics","level":3,"score":0.2565999925136566},{"id":"https://openalex.org/C195672273","wikidata":"https://www.wikidata.org/wiki/Q6888132","display_name":"Model-based design","level":2,"score":0.25279998779296875},{"id":"https://openalex.org/C150415221","wikidata":"https://www.wikidata.org/wiki/Q40687","display_name":"Robotic arm","level":2,"score":0.2502000033855438}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/iros60139.2025.11247019","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iros60139.2025.11247019","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320334627","display_name":"Engineering and Physical Sciences Research Council","ror":"https://ror.org/0439y7842"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":20,"referenced_works":["https://openalex.org/W3017024317","https://openalex.org/W3040646053","https://openalex.org/W3203385474","https://openalex.org/W3217214923","https://openalex.org/W4317934857","https://openalex.org/W4389402106","https://openalex.org/W4389518864","https://openalex.org/W4389520478","https://openalex.org/W4393235577","https://openalex.org/W4393578753","https://openalex.org/W4396694994","https://openalex.org/W4399767172","https://openalex.org/W4401414394","https://openalex.org/W4402330718","https://openalex.org/W4403919220","https://openalex.org/W4405785122","https://openalex.org/W4405787555","https://openalex.org/W4405846512","https://openalex.org/W4405909140","https://openalex.org/W4405910844"],"related_works":[],"abstract_inverted_index":{"Integrating":[0],"Large":[1],"Language":[2],"Models":[3],"(LLMs)":[4],"into":[5],"modern":[6],"robotic":[7,19,76],"systems":[8],"presents":[9,23],"significant":[10],"computational":[11,43,95,156],"and":[12,59,96,109,125,133],"energy":[13,97],"constraint":[14],"challenges,":[15],"particularly":[16],"for":[17,29,104],"human-centered":[18],"applications.":[20],"This":[21],"paper":[22],"a":[24,68,73,152],"novel":[25],"hardware":[26],"optimization":[27,86],"technique":[28],"deploying":[30,67],"LLMs":[31],"on":[32,72],"resource-constrained":[33],"embedded":[34,54],"devices,":[35],"achieving":[36],"an":[37,46],"up":[38,122],"to":[39,51,123,130],"77%":[40],"reduction":[41],"in":[42,49,107,155],"latency":[44],"through":[45],"FPGA":[47,82,144,149],"implementation":[48,120],"comparison":[50],"other":[52],"popular":[53],"computing":[55],"devices":[56],"(e.g.,":[57],"CPU":[58,132],"GPUs).":[60],"Additionally,":[61],"we":[62],"demonstrate":[63,114],"our":[64],"methodology":[65],"by":[66],"LLaMA":[69,118],"2-7B":[70,119],"model":[71],"Unitree":[74],"Go2":[75],"dog":[77],"integrated":[78],"with":[79],"the":[80,116,142],"proposed":[81,85,143],"platform.":[83],"The":[84],"framework":[87],"preserves":[88],"real-time":[89],"interaction":[90,106],"capabilities":[91],"while":[92,136],"significantly":[93],"reducing":[94],"overhead,":[98],"facilitating":[99],"efficient":[100],"natural":[101],"language":[102],"processing":[103],"human-robot":[105],"safety-critical":[108],"dynamic":[110],"environments.":[111],"Experimental":[112],"results":[113],"that":[115],"FPGA-based":[117],"achieves":[121],"6.06-fold":[124],"1.95-fold":[126],"higher":[127],"throughput":[128],"compared":[129],"baseline":[131],"GPU":[134],"implementations":[135],"maintaining":[137],"comparable":[138],"inference":[139],"accuracy.":[140],"Furthermore,":[141],"design":[145],"surpasses":[146],"existing":[147],"state-of-the-art":[148],"implementations,":[150],"delivering":[151],"30%":[153],"improvement":[154],"efficiency.":[157]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-11-28T00:00:00"}