Amazon Technologies, Inc.

Varadarajan Gopalakrishnan

Subram Narasimhan

In Chul Hyun

Jesper Mikael Johansson

Brandon William Porter

Andrew Jay Roths

Jin Dong Kim

Jon Arron McClintock

Darren Ernest Canavor

Kiran Kumar Edara

Cheol Su Kim

Andrew David Price

Dilip Kumar

Sridhar Boyapati

Siva Rama Kumar Somayajula

Omar FawazHashim Zakaria

Omar Fawazhashim Zakaria

Tzung-I Lee

Clifton Liu

Yi Zhu

James Domit Mackraz

Sheung Mei Chan

Shih-Chun Chang

David C Buuck

Kenneth Paul Kiraly

John Mathew Depew

Edward James Gayles

Gregg Elliott Zehr

Jeromey Russell Goetz

James Samuel Bowen

Bhavnish H Lathia

Jungtao Liu

Girish Bansilal Bajaj

Jay Praful Desai

Christopher Shingee Park

Lori Yoshida

David Y Woo

Albert William Wegener

Xi Chelsea Chen

Daniel Buchmueller

Avinash Joshi

Ilya Vladimirovich Brailovskiy

Kun Ting Tsai

Fabian Hensel

Morris Yuanhsiang Hsu

Huanchun Ye

QingYun Wei

Adam Barry Fineberg

Kwok Shing Lee

Kevin Michael Small

James Edwin Christy

Sang Eun Lee

Chirag Saurabh Bhavsar

Robert D Lau

Omar Fawaz Hashim Zakaria

This disclosure describes a system for managing inventory as it transitions into a materials handling facility, as it transitions between locations within a materials handling facility and/or as it transitions out of a materials handling facility. In some instances, a user (e.g., picker or picking agent) may retrieve an item from an inventory location and place the item into a tote. The systems described herein detect the item when it is added to or removed from the tote.

Image-based item detection and allocation

This disclosure describes a distributed automated mobile vehicle ('automated mobile vehicle') system for autonomously delivering orders of items to various delivery locations and/or autonomously returning items to a return location. In some implementations, each user may own or be assigned their own automated mobile vehicle that is associated with the user and an automated mobile vehicle control system maintained by the user. When the user orders an item, the user owned or controlled automated mobile vehicle navigates to a materials handling facility, retrieves the ordered item and delivers it to the user.

Distributed automated mobile vehicle routing based on characteristic information satisfying a minimum requirement

Item detection and transitions

To reduce the reliance on software for complex computations used in machine sensory perception, a sensory perception accelerator may include a neural network accelerator a linear algebra accelerator. The neural network accelerator may include systolic arrays to perform neural network computation circuits concurrently on image data and audio data. The linear algebra accelerator may include matrix computation circuits operable to perform matrix operations on image data and motion data.

Sensory perception accelerator

Techniques are generally described for network pattern matching. In various examples, first data may be sent over a network at a first bit rate to a second device. A plurality of network congestion profiles of the network may be stored in a memory. Network conditions of the network may be determined over a first period of time. A correlation between the network conditions and a first network congestion profile of the plurality of network congestion profiles may be determined. A second bit rate may be determined based on the first network congestion profile. Second data may be sent over the network to the second device at the second bit rate.

Network pattern matching

Wireless mesh network (WMN) architectures of network hardware devices organized in a mesh topology is described. One device communicates, using a first radio, first data with a second device via a first wireless link between the device and the second device. The device communicates, using a second radio, second data with a third device via a second wireless link between the device and the third device. The device communicates, using a third radio, third data with a fourth device via a third wireless link between the device and the fourth device. The device communicates, using a fourth radio, fourth data with a server of a content delivery network (CDN) via a point-to-point wireless link between the device and the server. The device is an only ingress point for content files for a mesh network that includes at least the device, the second device, and the third device.

Network hardware devices organized in a wireless mesh network for content distribution to client device having no internet connectivity

Inventory transitions

Two unmanned vehicles come within communication range of one another. The unmanned vehicles exchange logs of messages each has received. Each of the unmanned vehicles analyzes the messages that it received from the other unmanned vehicle to determine whether any of the received messages warrants changing a set of tasks it was planning to perform. When a message indicates that a task should be changed, the task is updated accordingly.

Unmanned vehicle message exchange

Tote based item tracking

Inventory tracking

A mesh network device includes a content repository including a first section to store original data and a second section to store cached data, and an application processor including a user space, kernel space, and an on-path data caching engine. The on-path data caching engine is to receive, from a second mesh network device, data frames via a data link layer of a TCP/IP mesh network. The data frames include audio or video content. The on-path data caching engine is further to store the data frames in a socket kernel buffer, determine that a third mesh network device is a destination for the data frames, and forward the data frames to the third mesh network device. The on-path data caching engine is to determine that the audio or video content is to be cached and asynchronously copy the audio or video content to the second section of the content repository.

On-path data caching in a mesh network

Distributed automated mobile vehicle routing

Described are techniques for determining occurrence of an error within a system of related devices. If the device affected by the error is not known, other devices within the system may be used to determine whether a device is unresponsive, non-functional, or limited in functionality. The other devices may also be used to determine data indicative of characteristics of the error, such as statuses and activities of devices at or near the time of the error, which may be used to generate a query. The query may be used to determine possible control actions that may address the error or prevent subsequent errors.

System for determining errors associated with devices

Described are techniques for determining occurrence of an error associated with a user device and user interfaces to present to users responsive to the error to acquire additional information. Particular user interfaces that correspond to an error may be determined based on the characteristics of the error. Particular communication channels by which the user interfaces may be provided to a user may be determined based on the devices affected by the error, the characteristics of the user interfaces, and the characteristics of the receiving devices. User input received responsive to a user interface may be associated with a particular error or user device and used to determine an action to be performed responsive to the error.

System for determining error data

Wireless mesh network (WMN) architectures of network hardware devices organized in a mesh topology in which the network hardware devices cooperate in distribution of content files to client consumption devices in an environment of limited connectivity to broadband Internet infrastructure is described. A self-contained, fully connected WMN can be used for localized delivery of content files. One WMN includes a single ingress node for ingress of content files into the wireless mesh network. The WMN also includes multiple network hardware devices wirelessly connected through a network backbone formed by multiple P2P wireless connections. A first network hardware device is wirelessly connected to a client consumption device by a first node-to-client (N2C) wireless connection and a second network hardware device is wirelessly connected to the single ingress node.

Network hardware devices organized in a wireless mesh network for content distribution to client devices having no internet connectivity

Growing community of inventors

Cupertino, CA, United States of America

Varadarajan Gopalakrishnan