| 1. |
- Bouyoucef, S E, et al.
(författare)
-
Poster Session 2 : Monday 4 May 2015, 08
- 2015
-
Ingår i: European Heart Journal Cardiovascular Imaging. - : Oxford University Press (OUP). - 2047-2404 .- 2047-2412. ; 16 Suppl 1
-
Tidskriftsartikel (refereegranskat)
|
|
| 2. |
- Bodin, Kenneth, et al.
(författare)
-
Constraint based particle fluids on GPGPU
- 2011
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- We present a fluid simulation method adapted for stream parallelism on general purpose graphics processingunits (GPGPU). In this method the equations of Navier and Stokes are discretized using particles and kernelfunctions as in Smoothed Particle Hydrodynamics (SPH), but rather than using penalty methods or solving for a divergence free velocity field, incompressibility is enforced using holonomic kinematic constraints [1]. We useone constraint for each smoothed particle stating that the local density should be kept constant. Other constraintsare used for boundary conditions and multiphysics coupling. We also present a viscosity model in which theshear rate at each pseudo particle is constrained to satisfy a given constitutive law. The computation of theconstraint forces, namely, the pressure and the stresses, requires the solution system of linear equations whichhave a sparse, saddle point structure. These are solved using the Uzawa method of preconditioned conjugate gradients (CG) applied directly to the symmetric indefinite matrix. The overall simulation method has its rootsin a discrete variational principle and the SPOOK time stepping scheme for constrained mechanical systems [2].The SPOOK method is second order accurate on the positions and constraints violations, and is stable at largetime-steps, thus often allowing several orders of magnitude larger timesteps in our method compared to intraditional SPH methods. The numerical implementation on GPGPU that is the main result of this paper consistsof the following components: particle neighbour searches based on spatial decomposition; summation of kernel densities; construction of Jacobians representing the constraints on the density, boundary conditions, viscosityand multiphysics couplings; a Uzawa CG solver for the system of linear equations; and finally, discrete timestepping of velocities and positions. The CG solver is particularly suitable for stream computing since it is basedon matrix-vector multiplications. The sparse system data is stored in a compressed matrix format and the algorithms operating on this data on GPGPU are implemented in CUDA and OpenCL. Our simulation resultsinclude performance measurements, and validation of the method for benchmark problems. We achieve up totwo orders of magnitude speed-up from the GPGPU over traditional processors and together with the increased timestep efficiency of our method we arrive at interactive performance for systems with up to two million fluidparticles representing an incompressible fluid.
|
|
| 3. |
- Andersson, Jennifer, et al.
(författare)
-
Predicting Gripability Heatmaps using Conditional GANs
- 2021
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- The feasibility of using conditional GANs (Generative Adversarial Networks) to predict gripability in log piles is investigated. This is done by posing gripability heatmap prediction from RGB-D data as an image-to-image translation problem. Conditional GANs have previously achieved impressive results on several image-to-image translation tasks predicting physical properties and adding details not present in the input images. Here, piles of logs modelled as sticks or rods are generated in simulation, and groundtruth gripability maps are created using a simple algorithm streamlining the datacollection process. A modified SSIM (Structural Similarity Index) is used to evaluate the quality of the gripability heatmap predictions. The results indicate promising model performance on several different datasets and heatmap designs, including using base plane textures from a real forest production site to add realistic noise in the RGB data. Including a depth channel in the input data is shown to increase performance compared to using pure RGB data. The implementation is based on the general Pix2Pix network developed by Isola et al. in 2017. However, there is potential to increase performance and model generalization, and the adoption of more advanced loss functions and network architectures are suggested. Next steps include using terrains reconstructed from highdensity laser scans in physics-based simulation for data generation. A more in-depth discussion regarding the level of sophistication required in the gripability heatmaps should also be carried out, along with discussions regarding other specifications that will be required for future deployment. This will enable derivation of a tailored gripability metric for ground-truth heatmap generation, and method evaluation on less ideal data.
|
|
| 4. |
- Andersson, Jennifer, et al.
(författare)
-
Reinforcement Learning Control of a Forestry Crane Manipulator
- 2021
-
Ingår i: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2021). - Prague : IEEE Robotics and Automation Society. - 9781665417150 - 9781665417143 ; , s. 2121-2126
-
Konferensbidrag (refereegranskat)abstract
- Forestry machines are heavy vehicles performing complex manipulation tasks in unstructured production forest environments. Together with the complex dynamics of the on-board hydraulically actuated cranes, the rough forest terrains have posed a particular challenge in forestry automation. In this study, the feasibility of applying reinforcement learning control to forestry crane manipulators is investigated in a simulated environment. Our results show that it is possible to learn successful actuator-space control policies for energy efficient log grasping by invoking a simple curriculum in a deep reinforcement learning setup. Given the pose of the selected logs, our best control policy reaches a grasping success rate of 97%. Including an energy-optimization goal in the reward function, the energy consumption is significantly reduced compared to control policies learned without incentive for energy optimization, while the increase in cycle time is marginal. The energy-optimization effects can be observed in the overall smoother motion and acceleration profiles during crane manipulation.
|
|
| 5. |
- Aoshima, Koji, et al.
(författare)
-
Data-driven models for predicting the outcome of autonomous wheel loader operations
-
Annan publikation (övrigt vetenskapligt/konstnärligt)abstract
- This paper presents a method using data-driven models for selecting actions and predicting the total performance of autonomous wheel loader operations over many loading cycles in a changing environment. The performance includes loaded mass, loading time, work. The data-driven models input the control parameters of a loading action and the heightmap of the initial pile state to output the inference of either the performance or the resulting pile state. By iteratively utilizing the resulting pile state as the initial pile state for consecutive predictions, the prediction method enables long-horizon forecasting. Deep neural networks were trained on data from over 10,000 random loading actions in gravel piles of different shapes using 3D multibody dynamics simulation. The models predict the performance and the resulting pile state with, on average, 95% accuracy in 1.2 ms, and 97% in 4.5 ms, respectively. The performance prediction was found to be even faster in exchange for accuracy by reducing the model size with the lower dimensional representation of the pile state using its slope and curvature. The feasibility of long-horizon predictions was confirmed with 40 sequential loading actions at a large pile. With the aid of a physics-based model, the pile state predictions are kept sufficiently accurate for longer-horizon use.
|
|
| 6. |
- Aoshima, Koji, et al.
(författare)
-
Examining the simulation-to-reality-gap of a wheel loader interacting with deformable terrain
- 2022
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- Simulators are essential for developing autonomous control of off-road vehicles and heavy equipment. They allow automatic testing under safe and controllable conditions, and the generation of large amounts of synthetic and annotated training data necessary for deep learning to be applied [1]. Limiting factors are the computational speed and how accurately the simulator reflects the real system. When the deviation is too large, a controller transfers poorly from the simulated to the real environment. On the other hand, a finely resolved simulator easily becomes too computationally intense and slow for running the necessary number of simulations or keeping realtime pace with hardware in the loop.We investigate how well a physics-based simulator can be made to match its physical counterpart, a full-scale wheel loader instrumented with motion and force sensors performing a bucket filling operation [2]. The simulated vehicle is represented as a rigid multibody system with nonsmooth contact and driveline dynamics. The terrain model combines descriptions of the frictional-cohesive soil as a continuous solid and particles, discretized in voxels and discrete elements [3]. Strong and stable force coupling with the equipment is mediated via rigid aggregate bodies capturing the bulk mechanics of the soil. The results include analysis of the agreement between a calibrated simulation model and the field tests, and of how the simulation performance and accuracy depend on spatial and temporal resolution. The system’s degrees of freedom range from hundreds to millions and the simulation speed up to ten times faster than realtime. Furthermore, it is investigated how sensitive a deep learning controller is to variations in the simulator environment parameters.[1] S. Backman, D. Lindmark, K. Bodin, M. Servin, J. Mörk, and H. Löfgren. Continuous control of an underground loader using deep reinforcement learning. Machines 9(10): 216 (2021).[2] K. Aoshima, M. Servin, E. Wadbro. Simulation-Based Optimization of High-Performance Wheel Loading. Proc. 38th Int. Symp. Automation and Robotics in Construction (ISARC), Dubai, UAE (2021).[3] M. Servin., T. Berglund., and S. Nystedt. A multiscale model of terrain dynamics for real-time earthmoving simulation. Advanced Modeling and Simulation in Engineering Sciences 8, 11 (2021).
|
|
| 7. |
- Aoshima, Koji, et al.
(författare)
-
Simulation-Based Optimization of High-Performance Wheel Loading
- 2021
-
Ingår i: Proceedings of the 38th International Symposium on Automation and Robotics in Construction. - Dubai : International Association for Automation and Robotics in Construction (IAARC). - 9789526952413 ; , s. 688-695
-
Konferensbidrag (refereegranskat)abstract
- Having smart and autonomous earthmoving in mind, we explore high-performance wheel loading in a simulated environment. This paper introduces a wheel loader simulator that combines contacting 3D multibody dynamics with a hybrid continuum-particle terrain model, supporting realistic digging forces and soil displacements at real-time performance. A total of 270,000 simulations are run with different loading actions, pile slopes, and soil to analyze how they affect the loading performance. The results suggest that the preferred digging actions should preserve and exploit a steep pile slope. High digging speed favors high productivity, while energy-efficient loading requires a lower dig speed.
|
|
| 8. |
- Aoshima, Koji, et al.
(författare)
-
World modeling for autonomous wheel loaders
- 2024
-
Ingår i: Automation. - : MDPI. - 2673-4052. ; 5:3, s. 259-281
-
Tidskriftsartikel (refereegranskat)abstract
- This paper presents a method for learning world models for wheel loaders performing automatic loading actions on a pile of soil. Data-driven models were learned to output the resulting pile state, loaded mass, time, and work for a single loading cycle given inputs that include a heightmap of the initial pile shape and action parameters for an automatic bucket-filling controller. Long-horizon planning of sequential loading in a dynamically changing environment is thus enabled as repeated model inference. The models, consisting of deep neural networks, were trained on data from a 3D multibody dynamics simulation of over 10,000 random loading actions in gravel piles of different shapes. The accuracy and inference time for predicting the loading performance and the resulting pile state were, on average, 95% in 1.21.2 ms and 97% in 4.54.5 ms, respectively. Long-horizon predictions were found feasible over 40 sequential loading actions.
|
|
| 9. |
|
|
| 10. |
- Backman, Sofi, et al.
(författare)
-
Continuous Control of an Underground Loader Using Deep Reinforcement Learning
- 2021
-
Ingår i: Machines. - : MDPI. - 2075-1702. ; 9:10
-
Tidskriftsartikel (refereegranskat)abstract
- The reinforcement learning control of an underground loader was investigated in a simulated environment by using a multi-agent deep neural network approach. At the start of each loading cycle, one agent selects the dig position from a depth camera image of a pile of fragmented rock. A second agent is responsible for continuous control of the vehicle, with the goal of filling the bucket at the selected loading point while avoiding collisions, getting stuck, or losing ground traction. This relies on motion and force sensors, as well as on a camera and lidar. Using a soft actor–critic algorithm, the agents learn policies for efficient bucket filling over many subsequent loading cycles, with a clear ability to adapt to the changing environment. The best results—on average, 75% of the max capacity—were obtained when including a penalty for energy usage in the reward.
|
|
