|
Position Calibration Method for Large size Industrial Robots Based on Random Forest |
|
Daiki Kato, Naoki Maeda, Toshiki Hirogaki, Eiichi Aoyama(Doshisha University, Japan), Kenichi Takahashi(IHI Corporation, Japan) |
| Most industrial robots are unsuitable for variable production systems because they are taught using the teaching playback method. In contrast, the offline teaching method has been developed, but it has not progressed because of the low positioning accuracy. We applied the random forest method and constructed a prediction model for positioning errors. The model to predict the positioning error from end-effector coordinates, joint angles, and joint torques was constructed, and the positioning error was predicted with high accuracy. The random forest analysis demonstrated that joint 2 was the primary factor. The positioning error norm was reduced at all points using the proposed calibration. |
|
|
Camouflaged Adversarial Attack on Object Detector |
|
Jeonghun Kim, Kyungmin Lee, Hyeongkeun Lee, Hunmin Yang, Se-Yoon Oh(Agency for Defense Development, Korea) |
| The existence of physical-world adversarial examples proves the vulnerability of deep learning systems. Therefore, it is essential to develop efficient adversarial attack algorithms to identify potential risks and build a robust system. The patch-based physical adversarial attack is effective way against neural network-based object detectors. However, the generated patches are quite perceptible for humans, violating the fundamental assumption of adversarial examples. In this work, we present task-specific loss functions that can generate imperceptible adversarial patches based on camouflaged patterns. |
|
|
LSTM-Based Real-Time SOC Estimation of Lithium-Ion Batteries Using a Vehicle Driving Simulator |
|
Si Jin Kim, Jong Hyun Lee, Dong Hun Wang, In Soo Lee(Kyungpook National University, Korea) |
| Currently, lithium-ion batteries (a type of secondary battery) are used as the primary sources of power in many
applications due to their low energy loss as a result of their high energy density and low self-discharge rate, and their ability to store energy for a long time. However, due to the frequent charging and discharging of such batteries,
overcharging is inevitable. This can cause system shutdowns, accidents, or property damage due to explosions. Therefore,
it is necessary to accurately predict the state of charge (SOC) of batteries for stable and efficient usage. |
|
|
Autonomous System Identification and Control Using Deep Neural Network |
|
Amirhosein Ghasemabadi, Benyamin Mehmandar, Ahmad Kalhor(University of Tehran, Iran, Islamic Republic of) |
| This paper aims to develop a parametric identification design for stable minimum-phase linear time-invariant(LTI) systems using Deep Learning. The proposed method employs Deep Learning and cosine similarity to identify the system’s type and parameters, which are then passed to a Sliding Mode Controller (SMC) to control the system. The presented approach achieves high performance in the identification and control process, requiring only a few system stimulation response samples to infer the system’s parameters. The results show our method can identify the type of systems with high accuracy and control them robustly independent of the slight error in system estimation. |
|
|
Feature Recognition for Graph-based Assembly Product Representation Using Machine Learning |
|
Jonathan Wörner, Daniella Brovkina, Oliver Riedel(University of Stuttgart, Germany) |
| The automation of the entire value chain of a product requires not only automated production and assembly, but also automated production and assembly planning. Today, assembly planning is mostly a manual task and is based on computer-aided design descriptions and technical drawings. For the automation of the assembly planning, an automated
recognition of the assembly features of the goal product is essential. In this paper, a concept for an automatic recognition of form features as well as the assignment of the connections between them as joints to build the assembly feature is
described. The feature recognition is based on point clouds and uses the PointNet architecture. |
|
|
Development of Chatbot Psychologist Applying Natural Language Understanding Techniques |
|
Batyrkhan Omarov(Al-Farabi Kazakh National University, Kazakhstan), Sergazy Narynov, Zhandos Zhumanov, Aidana Gumar, Mariyam Khassanova(Alem Research, Kazakhstan) |
| The importance of maintaining mental health has been recognized around the world, but the number of professionals offering such services has increased slightly, and therefore not everyone has the opportunity to receive consultations on a regular basis. According to statistics, the majority of people experienced anxiety and stress and at the same time noted that they had never visited a psychologist or a psychotherapist. The overall level of satisfaction with the services received in hospital institutions was low. As one of the reasons, respondents noted a long waiting period for an appointment with a specialist. |
|