Abstract: The present disclosure provides a power unit, aiming to provide a four-in-one drive powertrain for an electric vehicle. Key points of the technical scheme include a shell body. One side of the shell body is formed with a chamber having a partition structure arranged in a middle part thereof, the partition structure separates the chamber into a control part and a drive part, and the other side of the shell body is provided with a deceleration part and a DC/DC part. The control part includes a containing cavity positioned on a left side of the chamber and integrally formed with the shell body to place a controller. The drive part includes an accommodating cavity positioned on a right side of the chamber and integrally formed with the shell body to install an electric motor.

Abstract: An in-situ dynamic settlement test bench and method for the foundation soil is provided. The in-situ dynamic settlement test bench includes a vibration table system, a sensor embedded in the foundation soil, and a water injection trench around the foundation soil. A horizontal vibration load is applied to the lower foundation soil at an in-situ testing site by means of a vibration table, and the dynamic response of the foundation soil is measured through the sensors embedded in different positions, to reflect the seismic settlement deformation of the soil body. Dynamic settlement tests on soils include indoor dynamic simple shear, triaxial and torsional shear tests. A method is proposed for measuring the dynamic response and settlement deformation of the foundation soil by using a vibrator's vibration action on an in-situ soil column. Measurements of the acceleration response and seismic settlement deformation are taken for the foundation soil under in-situ condition.

Abstract: The present disclosure provides a method and device for dynamic recognition of emotion based on facial muscle movement monitoring, including: obtaining muscle movement data on one side and continuous frame images on the other side; obtaining multiple emotional states related to emotions, associating each emotional state with its corresponding continuous frame images, associating each emotional state with the muscle movement data corresponding to the continuous frame images at the same location and at the same time, to form a training set, building an emotion recognition model; inputting the muscle movement data and continuous frame images obtained in real time into the emotion recognition model to obtain a corresponding emotional state. This method builds an emotion recognition model through muscle movement data and continuous frame image, and uses motion data to make up for the parts of small action images that cannot be collected, thereby obtaining accurate emotion detection results.