FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4367164444> ?p ?o ?g. }

• W4367164444 endingPage "70763" @default.
• W4367164444 startingPage "70741" @default.
• W4367164444 abstract "Visual Odometry (VO) systems are widely used to determine the position and orientation of a robot or camera in an unknown environment. They are deployed on resource-constrained platforms, such as drones and Virtual Reality (VR) or Augmented Reality (AR) headsets. VO systems harnessing modern System-on-Chip (SoCs) with integrated Field Programmable Gate Array (FPGA) have the potential to improve the overall systems performance. This paper explores the FPGA acceleration of sparse VO kernels using High-level Synthesis (HLS) as this kind of VO system has been designed to use with low-power SoCs. We show that both computational and data transfer overheads between the processing cores of the CPU of the SoC and the accelerators on the FPGA need to be optimized to obtain better end-to-end performance. This is a result of the additional data movement incurred when using an FPGA accelerator and also because of the sparse computational nature with predictable or random memory access patterns of the kernels involved. However, state-of-the-art HLS tools are not yet able to perform the required optimizations automatically because they usually assume that the kernels to be accelerated have dense computational patterns with regular memory access. In this paper we propose three, potentially generic, methods to reduce the data transfer between the CPU and the customised hardware kernels on the FPGA; these methods are: (a) approximation based on domain-specific knowledge, (b) image compression, and (c) the use of on-the-fly computation. We present a case study of the use of these methods on SVO, a state-of-the-art sparse VO system with a semi-direct front-end. We demonstrate that our proposed methods can reduce data transfer overhead to achieve better end-to-end performance and that they can be applied not only when using standard Xilinx HLS tools but also with other state-of-the-art HLS tools, such as HeteroFlow. Compared to the baseline performance of the original SVO software on an Arm CPU, our proposed methods assist the HLS and HeteroFlow designs to achieve a speedup of 2.4x and 2.14x, respectively, without noticeable accuracy loss. The HLS and HeteroFlow designs also achieve a 1.85x and 1.89x, respectively, improvement in energy efficiency on the SoC system used. Compared to the SVO software baseline running on the Intel Xeon CPU, our proposed methods assist the HLS and HeteroFlow designs to achieve 8.2x and 8.3x improvement in energy efficiency, respectively." @default.
• W4367164444 created "2023-04-28" @default.
• W4367164444 creator A5021096884 @default.
• W4367164444 creator A5026069638 @default.
• W4367164444 creator A5060763068 @default.
• W4367164444 creator A5087600309 @default.
• W4367164444 date "2023-01-01" @default.
• W4367164444 modified "2023-10-09" @default.
• W4367164444 title "Exploring Sparse Visual Odometry Acceleration With High-Level Synthesis" @default.
• W4367164444 cites W1491719799 @default.
• W4367164444 cites W1578285471 @default.
• W4367164444 cites W1755205674 @default.
• W4367164444 cites W1967549803 @default.
• W4367164444 cites W1968257358 @default.
• W4367164444 cites W1970504153 @default.
• W4367164444 cites W2000880458 @default.
• W4367164444 cites W2004316855 @default.
• W4367164444 cites W2047120166 @default.
• W4367164444 cites W2070923961 @default.
• W4367164444 cites W2087070363 @default.
• W4367164444 cites W2088202621 @default.
• W4367164444 cites W2124313187 @default.
• W4367164444 cites W2151103935 @default.
• W4367164444 cites W2154555084 @default.
• W4367164444 cites W2166029537 @default.
• W4367164444 cites W2274994580 @default.
• W4367164444 cites W2343695530 @default.
• W4367164444 cites W2754957417 @default.
• W4367164444 cites W2883702102 @default.
• W4367164444 cites W2909863283 @default.
• W4367164444 cites W2914411014 @default.
• W4367164444 cites W2946669096 @default.
• W4367164444 cites W2971714504 @default.
• W4367164444 cites W2996485093 @default.
• W4367164444 cites W3034421981 @default.
• W4367164444 cites W3035708866 @default.
• W4367164444 cites W3163071374 @default.
• W4367164444 cites W3216154510 @default.
• W4367164444 cites W3216184190 @default.
• W4367164444 cites W4211118386 @default.
• W4367164444 cites W4226157514 @default.
• W4367164444 cites W4254910091 @default.
• W4367164444 cites W4286249767 @default.
• W4367164444 cites W4286377498 @default.
• W4367164444 cites W4302564868 @default.
• W4367164444 cites W4308650867 @default.
• W4367164444 cites W4311839611 @default.
• W4367164444 cites W4313155565 @default.
• W4367164444 cites W4321637065 @default.
• W4367164444 doi "https://doi.org/10.1109/access.2023.3268992" @default.
• W4367164444 hasPublicationYear "2023" @default.
• W4367164444 type Work @default.
• W4367164444 citedByCount "0" @default.
• W4367164444 crossrefType "journal-article" @default.
• W4367164444 hasAuthorship W4367164444A5021096884 @default.
• W4367164444 hasAuthorship W4367164444A5026069638 @default.
• W4367164444 hasAuthorship W4367164444A5060763068 @default.
• W4367164444 hasAuthorship W4367164444A5087600309 @default.
• W4367164444 hasBestOaLocation W43671644441 @default.
• W4367164444 hasConcept C117896860 @default.
• W4367164444 hasConcept C121332964 @default.
• W4367164444 hasConcept C13164978 @default.
• W4367164444 hasConcept C149635348 @default.
• W4367164444 hasConcept C154945302 @default.
• W4367164444 hasConcept C41008148 @default.
• W4367164444 hasConcept C42935608 @default.
• W4367164444 hasConcept C74650414 @default.
• W4367164444 hasConcept C9390403 @default.
• W4367164444 hasConceptScore W4367164444C117896860 @default.
• W4367164444 hasConceptScore W4367164444C121332964 @default.
• W4367164444 hasConceptScore W4367164444C13164978 @default.
• W4367164444 hasConceptScore W4367164444C149635348 @default.
• W4367164444 hasConceptScore W4367164444C154945302 @default.
• W4367164444 hasConceptScore W4367164444C41008148 @default.
• W4367164444 hasConceptScore W4367164444C42935608 @default.
• W4367164444 hasConceptScore W4367164444C74650414 @default.
• W4367164444 hasConceptScore W4367164444C9390403 @default.
• W4367164444 hasFunder F4320316514 @default.
• W4367164444 hasFunder F4320320005 @default.
• W4367164444 hasFunder F4320320006 @default.
• W4367164444 hasFunder F4320334627 @default.
• W4367164444 hasLocation W43671644441 @default.
• W4367164444 hasOpenAccess W4367164444 @default.
• W4367164444 hasPrimaryLocation W43671644441 @default.
• W4367164444 hasRelatedWork W2160069347 @default.
• W4367164444 hasRelatedWork W2351404747 @default.
• W4367164444 hasRelatedWork W2373066471 @default.
• W4367164444 hasRelatedWork W2521029662 @default.
• W4367164444 hasRelatedWork W2531676558 @default.
• W4367164444 hasRelatedWork W2759844239 @default.
• W4367164444 hasRelatedWork W3011776263 @default.
• W4367164444 hasRelatedWork W3094426418 @default.
• W4367164444 hasRelatedWork W4300585894 @default.
• W4367164444 hasRelatedWork W4363649491 @default.
• W4367164444 hasVolume "11" @default.
• W4367164444 isParatext "false" @default.
• W4367164444 isRetracted "false" @default.
• W4367164444 workType "article" @default.