A fully integrated system for hardware-accelerated TSDF SLAM with LiDAR sensors (HATSDF SLAM)

Eisoldt, Marc; Gaal, Julian; Wiemann, Thomas; Flottmann, Marcel; Rothmann, Marc; Tassemeier, Marco; Porrmann, Mario

doi:10.1016/j.robot.2022.104205

Cited by 6 publications

(2 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other advances toward VO/SLAM acceleration include LiDAR SLAM accelerators in [36], [37], and [38], as well as accelerators for Iterative Closest Point (ICP) ( [34] and [35]).…”

Section: ) Accelerating Vo/slam Systems On Fpgasmentioning

confidence: 99%

Exploring Sparse Visual Odometry Acceleration With High-Level Synthesis

Iordanou

Riley

et al. 2023

IEEE Access

View full text Add to dashboard Cite

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 lowpower 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.

show abstract

“…Other advances toward VO/SLAM acceleration include LiDAR SLAM accelerators in [36], [37], and [38], as well as accelerators for Iterative Closest Point (ICP) ( [34] and [35]).…”

Section: ) Accelerating Vo/slam Systems On Fpgasmentioning

confidence: 99%

Exploring Sparse Visual Odometry Acceleration With High-Level Synthesis

Iordanou

Riley

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Lidar overcomes the shortcoming of traditional sensors in obtaining spatial information and can work all day long to obtain the point cloud data of real trench spatial information. Lidar has been widely used in obstacle detection (Qin et al, 2023;Shang et al, 2023), terrain mapping (Kim and Choi, 2021;García-López et al, 2023), map construction (Su et al, 2021;Ao et al, 2022;Eisoldt et al, 2022;Rivera et al, 2023), agricultural information monitoring, and plant model reconstruction (Perez et al, 2018;Tsoulias et al, 2019;Campbell et al, 2023).…”

Section: Introductionmentioning

confidence: 99%

Three-dimensional reconstruction of the furrow shape in orchards using a low-cost lidar

Zhou

Meng

et al. 2023

Front. Sustain. Food Syst.

View full text Add to dashboard Cite

Efficient furrow fertilization is extremely critical for fertilizer utilization, fruit yield, and fruit quality. The precise determination of trench quality necessitates the accurate measurement of its characteristic parameters, including its shape and three-dimensional structure. Some existing algorithms are limited to detecting only the furrow depth while precluding the tridimensional reconstruction of the trench shape. In this study, a novel method was proposed for three-dimensional trench shape reconstruction and its parameter detection. Initially, a low-cost multi-source data acquisition system with the 3D data construction method of the trench was developed to address the shortcomings of single-sensor and manual measurement methods in trench reconstruction. Subsequently, the analysis of the original point cloud clarified the “coarse-fine” two-stage point cloud filtering process, and then a point cloud preprocessing method was proposed based on ROI region extraction and discrete point filtering. Furthermore, by analyzing the characteristics of the point cloud, a random point preselection condition based on the variance threshold was designed to optimize the extraction method of furrow side ground based on RANSAC. Finally, a method was established for extracting key characteristic parameters of the trench and trench reconstruction based on the fitted ground model of the trench side. Experimental results demonstrated that the point cloud pretreatment method could eliminate 83.8% of invalid point clouds and reduce the influence of noise points on the reconstruction accuracy. Compared with the adverse phenomena of fitting ground incline and height deviation of the original algorithm results, the ground height fitted by the improved ditch surface extraction algorithm was closer to the real ground, and the identification accuracy of inner points of the ground point cloud was higher than that of the former. The error range, mean value error, standard deviation error, and stability coefficient error of the calculated ditch width were 0 ~ 5.965%, 0.002 m, 0.011 m, and 0.37%, respectively. The above parameters of the calculated depth were 0 ~ 4.54%, 0.003 m, 0.017 m, and 0.47%, respectively. The results of this research can provide support for the comprehensive evaluation of the quality of the ditching operation, the optimization of the structure of the soil touching part, and the real-time control of operation parameters.

show abstract