2results about How to "Delayed decay" patented technology

This utility model discloses a railing device, relating to the field of railing technology. The device includes a post, a crossbar, and a wooden wedge-shaped locking strip. The post has a rectangular groove on its side; the crossbar has an insertion part at its end, with a wedge-shaped groove at its top; the locking strip has a right-angled trapezoidal cross-section and achieves mechanical self-locking by inserting its inclined surface downwards into the wedge-shaped groove. When the locking strip is inserted, its inclined surface presses against the inner wall of the wedge-shaped groove, forcing the insertion part to expand inwards and press against the inner wall of the rectangular groove, forming a horizontal compressive force; simultaneously, the top of the locking strip presses against the upper edge of the rectangular groove, resisting vertical pull-out force. This device uses an all-wood structure, eliminating the risk of corrosion from metal nails and extending its service life; at the same time, the wedge-shaped self-locking structure provides strong resistance to pull-out and impact, preventing loosening, and is suitable for outdoor scenarios such as parks and boardwalks.

The invention discloses a micro light emitting diode dynamic control system, control method and control device based on a multi-core heterogeneous chip. A task management and light field analysis unit TMU writes an analyzed light field descriptor set into an on-chip shared memory and sends a doorbell signal to a neural mimicry computing cluster NPU; a neural mimicry computing cluster NPU receives the light field descriptor from a task management and light field analysis unit TMU, reads historical data from a distributed shared memory, writes a processing result into the distributed shared memory, and triggers a high-precision driving computing cluster HPU and a distributed pixel driving network DPS; the high-precision driving calculation cluster HPU receives the instruction and the data, reads accurate parameters from the distributed shared memory and issues a final driving micro instruction to the distributed pixel driving network DPS; and the distributed pixel driving network DPS is used for receiving microinstructions and local parameters from the high-precision driving calculation cluster HPU. And inherent contradictions among power consumption, bandwidth, service life and display quality are solved.