• Journal of Institute of Control, Robotics and Systems
• /
• v.7 no.9
• /
• pp.717-722
• /
• 2001

This paper presents a methodology on the backward path tracking control of a trailer type robot which consists of two parts: a tractor and a trailer. It is difficult to control the motion of a trailer vehicle since its dynamics is non-holonomic. Therefore, in this paper, the modeling and parameter estimation of the system using a real-coded genetic algorithm(RCGA) is proposed and a backward path tracking control algorithm is then obtained based on the linearized model. Experimental results verify the effectiveness of the proposed method.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.46 no.12
• /
• pp.108-116
• /
• 2009

In this paper, we propose a Ethernet based audio distribution system with a new L1/L2 hybrid switching scheme, and evaluate its performance. The proposed scheme not only offers guaranteed low latency and jitter characteristics that are essentially required for the distribution of high-quality uncompressed audio traffic, and but also provide an efficient transmission of data traffic on the Ethernet environment. The audio distribution system with a proposed scheme consists of a master node and a number of relay nodes, and all nodes are mutually connected as a daisy-chain topology through up and downlinks. The master node generates an audio frame for each cycle of 125us, and the audio frame has 24 time slotted audio channels for carrying stereo 24 channels of 16-bit PCM sampled audio. On receiving the audio frame from its upstream node via the downlink, each intermediate node inserts its audio traffic to the reserved time slot for itself, then relays again to next node through its physical layer(L1) transmission - repeating. After reaching the end node, the audio frame is loopbacked through the uplink. On repeating through the uplink, each node makes a copy of audio slot that node has to receive, then play the audio. When the audio transmission is completed, each node works as a normal L2 switch, thus data frames are switched during the remaining period. For supporting this L1/L2 hybrid switching capability, we insert a glue logic for parsing and multiplexing audio and data frames at MII(Media Independent Interlace) between the physical and data link layers. The proposed scheme can provide a good delay performance and transmission efficiency than legacy Ethernet based audio distribution systems. For verifying the feasibility of the proposed L1/L2 hybrid switching scheme, we use OMNeT++ as a simulation tool with various parameters. From the simulation results, one can find that the proposed scheme can provides outstanding characteristics in terms of both jitter characteristic for audio traffic and transmission efficiency of data traffics.

• Journal of Korean Society of Forest Science
• /
• v.110 no.4
• /
• pp.678-688
• /
• 2021

This study was conducted to provide growth characteristics and stand information in preparation for natural regeneration method and silvicultural treatment of seed tree method in Korean red pine stands by using the field inventoried data 18 years after the seed tree regeneration method in 2001 and analyzing the stand and sapling characteristics. The study area was located in the Research Forest of Kangwon National University, Chuncheon-si, Gangwon-do. In November 2018, we collected tree measurement factors, such as diameter at breast height (DBH), height, crown class, stem quality, tree age, and annual height increment on the established plots. We calculated the basal area, volume, height-diameter ratio (HD ratio), and additional stand density variables such as the relative density and %stocking to analyze the managed stands. The number of mother trees was 58 trees/ha, with a 17.7% stocking level, whereas the number of saplings was 2,330 trees/ha, with a 79.0% stocking level. In germination, the age distribution of saplings ranged from 2001 to 2007, with most belonging to 15-16 years. The development condition of saplings was 10 cm in DBH and 8 m in height. DBH and height were higher as the crown class is more dominant, and this difference was statistically significant in the analysis of variance and Duncan's multiple comparison test (p<0.0001). HD ratio ranged mostly between 80% and 90%, and more than 95% of sapling stems were of high quality, with a straight, unbroken top, non-sweep, and non-diseased stem. On average, the annual height increment of saplings was 21.9 cm at 1 year, 43.3 cm at 5 years, 54.3 cm at 10 years, and 64.3 cm at 15 years. The overall height growth with age increased smoothly. According to the analysis of covariance, the annual height growth by crown class differed significantly. The regression analysis parameters revealed that annual height growth increased with age and dominant crown class.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.163-177
• /
• 2019

As smartphones are getting widely used, human activity recognition (HAR) tasks for recognizing personal activities of smartphone users with multimodal data have been actively studied recently. The research area is expanding from the recognition of the simple body movement of an individual user to the recognition of low-level behavior and high-level behavior. However, HAR tasks for recognizing interaction behavior with other people, such as whether the user is accompanying or communicating with someone else, have gotten less attention so far. And previous research for recognizing interaction behavior has usually depended on audio, Bluetooth, and Wi-Fi sensors, which are vulnerable to privacy issues and require much time to collect enough data. Whereas physical sensors including accelerometer, magnetic field and gyroscope sensors are less vulnerable to privacy issues and can collect a large amount of data within a short time. In this paper, a method for detecting accompanying status based on deep learning model by only using multimodal physical sensor data, such as an accelerometer, magnetic field and gyroscope, was proposed. The accompanying status was defined as a redefinition of a part of the user interaction behavior, including whether the user is accompanying with an acquaintance at a close distance and the user is actively communicating with the acquaintance. A framework based on convolutional neural networks (CNN) and long short-term memory (LSTM) recurrent networks for classifying accompanying and conversation was proposed. First, a data preprocessing method which consists of time synchronization of multimodal data from different physical sensors, data normalization and sequence data generation was introduced. We applied the nearest interpolation to synchronize the time of collected data from different sensors. Normalization was performed for each x, y, z axis value of the sensor data, and the sequence data was generated according to the sliding window method. Then, the sequence data became the input for CNN, where feature maps representing local dependencies of the original sequence are extracted. The CNN consisted of 3 convolutional layers and did not have a pooling layer to maintain the temporal information of the sequence data. Next, LSTM recurrent networks received the feature maps, learned long-term dependencies from them and extracted features. The LSTM recurrent networks consisted of two layers, each with 128 cells. Finally, the extracted features were used for classification by softmax classifier. The loss function of the model was cross entropy function and the weights of the model were randomly initialized on a normal distribution with an average of 0 and a standard deviation of 0.1. The model was trained using adaptive moment estimation (ADAM) optimization algorithm and the mini batch size was set to 128. We applied dropout to input values of the LSTM recurrent networks to prevent overfitting. The initial learning rate was set to 0.001, and it decreased exponentially by 0.99 at the end of each epoch training. An Android smartphone application was developed and released to collect data. We collected smartphone data for a total of 18 subjects. Using the data, the model classified accompanying and conversation by 98.74% and 98.83% accuracy each. Both the F1 score and accuracy of the model were higher than the F1 score and accuracy of the majority vote classifier, support vector machine, and deep recurrent neural network. In the future research, we will focus on more rigorous multimodal sensor data synchronization methods that minimize the time stamp differences. In addition, we will further study transfer learning method that enables transfer of trained models tailored to the training data to the evaluation data that follows a different distribution. It is expected that a model capable of exhibiting robust recognition performance against changes in data that is not considered in the model learning stage will be obtained.