摘要 :
RFID-based human activity recognition(HAR)attracts attention due to its convenience,noninvasiveness,and privacy protection.Existing RFID-based HAR methods use modeling,CNN,or LSTM to extract features effectively.Still,they have sh...
展开
RFID-based human activity recognition(HAR)attracts attention due to its convenience,noninvasiveness,and privacy protection.Existing RFID-based HAR methods use modeling,CNN,or LSTM to extract features effectively.Still,they have shortcomings:1)requiring complex hand-crafted data cleaning processes and 2)only addressing single-person activity recognition based on specific RF signals.To solve these problems,this paper proposes a novel device-free method based on Time-streaming Multiscale Transformer called TransTM.This model leverages the Transformer''s powerful data fitting capabilities to take raw RFID RSSI data as input without pre-processing.Concretely,we propose a multiscale convolutional hybrid Transformer to capture behavioral features that recognizes singlehuman activities and human-to-human interactions.Compared with existing CNN-and LSTM-based methods,the Transformer-based method has more data fitting power,generalization,and scalability.Furthermore,using RF signals,our method achieves an excellent classification effect on human behaviorbased classification tasks.Experimental results on the actual RFID datasets show that this model achieves a high average recognition accuracy(99.1%).The dataset we collected for detecting RFID-based indoor human activities will be published.
收起
摘要 :
Several scenario experiments of the IPCC 4th Assessment Report (AR4) are performed by version g1.0 of a Flexible coupled Ocean-Atmosphere-Land System Model (FGOALS) developed at the Institute of At- mospheric Physics, Chinese Acad...
展开
Several scenario experiments of the IPCC 4th Assessment Report (AR4) are performed by version g1.0 of a Flexible coupled Ocean-Atmosphere-Land System Model (FGOALS) developed at the Institute of At- mospheric Physics, Chinese Academy of Sciences (IAP/CAS), including the "Climate of the 20th century experiment", "CO2 1% increase per year to doubling experiment" and two separate IPCC greenhouse gases emission scenarios AIB and B1 experiments. To distinguish between the different impacts of natural vari- ations and human activities on the climate change, three-member ensemble runs are performed for each scenario experiment. The coupled model simulations show: (1) from 1900 to 2000, the global mean temper- ature increases about 0.5℃ and the major increase occurs during the later half of the 20th century, which is in consistent with the observations that highlights the coupled model's ability to reproduce the climate changes since the industrial revolution; (2) the global mean surface air temperature increases about 1.6℃ in the CO2 doubling experiment and 1.5℃ and 2.4℃ in the AlB and Bl scenarios, respectively. The global warming is indicated by not only the changes of the surface temperature and precipitation but also the temperature increase in the deep ocean. The thermal expansion of the sea water would induce the rise of the global mean sea level. Both the control run and the 20th century climate change run are carried out again with version g1.1 of FGOALS, in which the cold biases in the high latitudes were removed. They are then compared with those from version g1.0 of FGOALS in order to distinguish the effect of the model biases on the simulation of global warming.
收起
摘要 :
This paper tries to qualitatively analyze land use and cover changes (LUCC) and accompanied soil degradation in China, the data of World Atlas of Agriculture in 1969, Land Cover Data of Asia in 1992 and Global Assessment of Human ...
展开
This paper tries to qualitatively analyze land use and cover changes (LUCC) and accompanied soil degradation in China, the data of World Atlas of Agriculture in 1969, Land Cover Data of Asia in 1992 and Global Assessment of Human Induced Soil Degradation in 1992, etc. have been used. From 1960' s to 1990's, the area of forestland had increased, arable land and paddy as well as grassland had significantly decreased in China. The major type of soil degradation is due to water erosion, which is widely distributed especially in Loess Plateau and in Southeast and Southwest China, this happened in forestland, grassland and arable land and mainly resulted from deforestation. The secondary type is wind erosion, which is mainly distributed in Northwest China, it happened in arable land and grassland and was caused by overgrazing. Chemical deterioration is distributed in North and Northwest China, which happened in paddy, arable land and grassland. Physical deterioration only concentrated in local area in North China, it happened in arable and paddy. Chemical and physical deteriora-tion mainly resulted from unreasonable activity. Because of different human activities, different LUCC and different natural conditions, the types and strength of soil degradation can be different.
收起
摘要 :
By using field salinity data recorded in the Qiantang Estuary for the nearly forty years (which covers the periods before and after the erection of a large-scale reservoir in the watershed and river regulation/reclamation in the e...
展开
By using field salinity data recorded in the Qiantang Estuary for the nearly forty years (which covers the periods before and after the erection of a large-scale reservoir in the watershed and river regulation/reclamation in the estuary) and one-dimensional salt water intrusion modeling, it is shown that the salt water intrusion decreases caused by increased low water discharge for the reservoir, while the intrusion increases for the increased tidal range because of the river regulation/reclamation. The integrate result is that the saltwater intrusion decreases.
收起
