US-12619229-B2 - Network aware and predictive motion planning in mobile multi-robotics systems

US12619229B2US 12619229 B2US12619229 B2US 12619229B2US-12619229-B2

Abstract

Techniques are disclosed to facilitate multi-agent path planning and to enable navigation for robotics systems to be more resilient to wireless network related issues. The discussed techniques include enhancing path-planning algorithms to consider wireless Quality of Service (QoS) metrics for the identification of planned multi-agent paths. Moreover, the techniques include the compensation of communication and computational latencies to enable offloading of time-sensitive navigation workloads to network infrastructure components.

Inventors

Susruth Sudhakaran
Gagan Acharya
Amit Baxi
Dave Cavalcanti
Mark Eisen
Ramya M
Javier Perez-Ramirez
Shagaya Mageshkumar Vincent

Assignees

INTEL CORPORATION

Dates

Publication Date: 20260505
Application Date: 20210324

Claims (20)

1 . A computing device, comprising: a communication interface to receive quality of service (QoS) metrics from an autonomous agent operating within an environment of a wireless network in which the autonomous agent communicates with the computing device via one or more wireless links, wherein the QoS metrics correspond to the one or more wireless links and are generated by the autonomous agent via monitoring of the one or more wireless links; and one or more processors configured to: generate an environment model comprising a plurality of grid cells of predetermined sizes, to which the QoS metrics are mapped, the environment model representing a model of the environment for navigation path planning of the autonomous agent within the environment; calculate, using the environment model, a route within the environment comprising a plurality of waypoints, the route being based upon a predicted network performance along the route with respect to communications between the autonomous agent and the computing device via the one or more wireless links for the autonomous agent when traversing the route; perform waypoint correction by (i) receiving a plurality of waypoints, (ii) predicting a trajectory delay of the autonomous agent caused by a communication delay and a computing delay, and (iii) calculating a plurality of delay-corrected waypoints, wherein the communication interface is configured to transmit, to the autonomous agent, a the plurality of delay-corrected waypoints as a command to cause the autonomous agent to traverse the calculated route by traversing the plurality of delay-corrected waypoints that comprise a subset of the plurality of grid cells of the environment model, wherein the plurality of delay-corrected waypoints that comprise the route within the environment that is based upon a predicted network performance with respect to communications between the autonomous agent and the computing device when the autonomous agent traverses the route, wherein the autonomous agent is from among a plurality of autonomous agents operating within the environment comprising autonomous mobile robots (AMRs), and wherein the environment model is shared among the plurality of autonomous agents for navigation path planning of each of the plurality of autonomous agents.
2 . The computing device of claim 1 , wherein the computing device is an Edge network computing device.
3 . The computing device of claim 1 , wherein the one or more processors are configured to aggregate the QoS metrics received from each of the plurality of autonomous agents to calculate an expected QoS score per grid cell that represents an expected QoS level in a respective grid cell.
4 . The computing device of claim 3 , wherein the one or more processors are configured to calculate the expected QoS score per grid cell using a weighted averaging function, with more recent values being assigned a higher weight.
5 . The computing device of claim 1 , wherein the calculated route comprises a chain of connected grid cells from among the plurality of grid cells that indicate a path within the environment for the autonomous agent to traverse to reach a destination grid cell.
6 . The computing device of claim 3 , wherein the one or more processors are configured to calculate the environment model that maps the expected QoS score per grid cell.
7 . The computing device of claim 1 , wherein the QoS metrics include: latency, receive signal strength indicator (RSSI) values, packet error rate (PER), jitter, bit error rate (BER), signal to noise ratio (SNR), signal to noise plus interference ratio (SINR), carrier to interference plus noise ratio (CINR), or modulation and coding schemes (MCS) histogram data.
8 . An autonomous agent, comprising: a communication interface to enable the autonomous agent to communicate with a computing device via a wireless network using one or more wireless links within an environment; and one or more processors configured to monitor the one or more wireless links and to generate quality of service (QoS) metrics with respect to the one or more wireless links from the monitoring, wherein the communication interface transmits the QoS metrics to the computing device for the generation of an environment model that is associated with the environment, wherein the environment model comprises a plurality of grid cells of predetermined sizes, to which the QoS metrics are mapped, wherein the one or more processors are configured to calculate, using the environment model, a route for the autonomous agent to use to navigate to a destination within the environment, wherein the communication interface is configured to receive, via the computing device, a plurality of delay-corrected waypoints as a command to cause the autonomous agent to traverse the calculated route by traversing the plurality of delay-corrected waypoints that comprise a subset of the plurality of grid cells of the environment model, wherein the plurality of delay-corrected waypoints are (i) computed based upon a predicted trajectory delay of the autonomous agent caused by a communication delay and a computing delay, and (ii) based upon a predicted network performance with respect to communications between the autonomous agent and the computing device when the autonomous agent traverses the route, wherein the calculated route is based upon a predicted network performance along the route with respect to communications between the autonomous agent and the computing device via the one or more wireless links for the autonomous agent when traversing the route, wherein the autonomous agent is from among a plurality of autonomous agents comprising autonomous mobile robots (AMRs) operating within the environment, and wherein the environment model is shared among the plurality of autonomous agents for navigation path planning of each of the plurality of autonomous agents.
9 . The autonomous agent of claim 8 , wherein the communication interface communicates with an Edge network computing device that comprises the computing device.
10 . The autonomous agent of claim 8 , wherein the environment model maps an expected QoS score per grid cell from an aggregation of QoS metrics transmitted by each of the plurality of autonomous agents.
11 . The autonomous agent of claim 10 , wherein the expected QoS score per grid cell is based upon a weighted averaging function, with more recent values being assigned a higher weight.
12 . The autonomous agent of claim 8 , wherein the calculated route comprises a chain of connected grid cells from among the plurality of grid cells that indicate a path within the environment for the autonomous agent to traverse to reach a destination grid cell corresponding to the destination within the environment.
13 . The autonomous agent of claim 10 , wherein the autonomous agent is configured to calculate the route by iteratively executing a cost function that uses the expected QoS score per grid cell and an estimate of a possible QoS score from a grid cell in the navigation route to reach a destination grid cell corresponding to the destination within the environment.
14 . The autonomous agent of claim 8 , wherein the QoS metrics include: latency, receive signal strength indicator (RSSI) values, packet error rate (PER), jitter, bit error rate (BER), signal to noise ratio (SNR), signal to noise plus interference ratio (SINR), carrier to interference plus noise ratio (CINR), or modulation and coding schemes (MCS) histogram data.
15 . A computing device, comprising: a communication interface to receive, via a wireless network, sensor data from an autonomous agent operating within an environment after the sensor data is transmitted by the autonomous agent in accordance with a communication delay; one or more processors configured to calculate, using an environment model comprising a plurality of grid cells of predetermined sizes, to which Quality of Service (QoS) metrics of the wireless network are mapped, a route for the autonomous agent to reach a destination within the environment using the sensor data, the route comprising a plurality of waypoints, and the calculation being associated with a computing delay; and a neural network trained to perform waypoint correction by (i) receiving the plurality of waypoints, (ii) predicting a trajectory delay of the autonomous agent caused by the communication delay and the computing delay, and (iii) calculating a plurality of delay-corrected waypoints, wherein the communication interface transmits the plurality of delay-corrected waypoints as a command to cause the autonomous agent to traverse a navigation path by traversing the plurality of delay-corrected waypoints that comprise a subset of the plurality of grid cells of the environment model, wherein the plurality of delay-corrected waypoints that comprise the route within the environment that is based upon a predicted network performance with respect to communications between the autonomous agent and the computing device for the autonomous agent when traversing the route, wherein the autonomous agent is from among a plurality of autonomous agents comprising autonomous mobile robots (AMRs) operating within the environment, and wherein the environment model is shared among the plurality of autonomous agents for navigation path planning of each of the plurality of autonomous agents.
16 . The computing device of claim 15 , wherein the computing device is an Edge network computing device.
17 . The computing device of claim 15 , wherein the neural network comprises a deep neural network implementing a long short-term memory (LSTM) architecture.
18 . The computing device of claim 15 , wherein the neural network calculates the plurality of delay-corrected waypoints further using a predicted delay associated with downlink communication latency with respect to a time required to transmit the plurality of delay-corrected waypoints to the autonomous agent.
19 . The computing device of claim 15 , wherein: the one or more processors are configured to calculate the route for the autonomous agent using the environment model having an expected QoS score per grid cell mapped to the plurality of grid cells, the expected QoS score being calculated using an aggregation of a respective set of QoS metrics received from each one of the plurality of autonomous agents, wherein the QoS metrics correspond to one or more wireless links used for communications between the computing device and the plurality of autonomous agents.
20 . The computing device of claim 19 , wherein the QoS metrics include: latency, receive signal strength indicator (RSSI) values, packet error rate (PER), jitter, bit error rate (BER), signal to noise ratio (SNR), signal to noise plus interference ratio (SINR), carrier to interference plus noise ratio (CINR), or modulation and coding schemes (MCS) histogram data.

Description

TECHNICAL FIELD Aspects described herein generally relate to motion planning for autonomous systems and, more particularly, to techniques implementing wireless quality of service (QoS) information and machine learning to provide motion planning for autonomous agents. BACKGROUND The upcoming “Industry 4.0” paradigm is using computing, communication, and AI technologies to increase automation efficiency, reduce energy use, and improve safety for human workers. Autonomous Mobile Robots (AMRs) are key components in factories and warehouses to meet these needs. In accordance with such use cases, AMRs implement perception and manipulation jointly to accomplish a given task by navigating an environment while communicating and coordinating with one other as well as with a central entity. This coordination requires bounded latency, reliable communications, and computing power, which is typically handled in a networked environment that services the AMR environment. However, current techniques to support AMR path planning and navigation in such environments have been inadequate. BRIEF DESCRIPTION OF THE DRAWINGS/FIGURES The accompanying drawings, which are incorporated herein and form a part of the specification, illustrate the aspects of the present disclosure and, together with the description, and further serve to explain the principles of the aspects and to enable a person skilled in the pertinent art to make and use the aspects. FIG. 1 illustrates a block diagram of an exemplary environment utilizing autonomous mobile robots (AMRs), in accordance with aspects of the disclosure. FIG. 2A illustrates a block diagram of multi-agent path planning process flow based on wireless Quality of Service (QoS) information, in accordance with aspects of the disclosure. FIG. 2B illustrates an exemplary graph showing the relationship between a latency QoS metric and a resulting QoS score, in accordance with aspects of the disclosure. FIG. 2C illustrates a block diagram of an exemplary shared environment model incorporating wireless QoS data, in accordance with aspects of the disclosure. FIG. 3 illustrates a block diagram of a process flow associated with offloading navigational tasks for a multi-agent navigation pipeline and performing waypoint correction, in accordance with aspects of the disclosure. FIG. 4 illustrates a block diagram of an exemplary autonomous agent, in accordance with an aspects of the disclosure. FIG. 5 illustrates a block diagram of an exemplary computing device, in accordance with an aspects of the disclosure. The exemplary aspects of the present disclosure will be described with reference to the accompanying drawings. The drawing in which an element first appears is typically indicated by the leftmost digit(s) in the corresponding reference number. DETAILED DESCRIPTION In the following description, numerous specific details are set forth in order to provide a thorough understanding of the aspects of the present disclosure. However, it will be apparent to those skilled in the art that the aspects, including structures, systems, and methods, may be practiced without these specific details. The description and representation herein are the common means used by those experienced or skilled in the art to most effectively convey the substance of their work to others skilled in the art. In other instances, well-known methods, procedures, components, and circuitry have not been described in detail to avoid unnecessarily obscuring aspects of the disclosure. Again, current techniques to current techniques to support AMR route planning and navigation in such environments have various drawbacks. For instance, networks implemented in AMR environments may include Wireless Time Sensitive Networking (WTSN) systems, such as infrastructure Wi-Fi and cellular-based network solutions that may be deployed in a warehouse or factory to facilitate AMR autonomous functions. However, one of the challenges associated with a WTSN infrastructure deployment is that the communication environment is highly dynamic and difficult to address with wireless planning alone. For instance, an AMR environment such as a factory or warehouse may use a wireless network that provides inconsistent levels of wireless communication performance due to changes in the environment, such as AMRs or other objects moving through the wireless communication medium. The wireless communication performance may also suffer from inconsistencies due to these changes over time, which may be characterized by increased latency, jitter, etc. due to factors such as cross channel interference, shielding, roaming, etc. Thus, in such use cases there exists a need to ensure a consistent level of connectivity, which is characterized by bounded latency and reliability. Conventionally, a multi-robot path planning algorithm may be implemented for planning paths for all AMRs in a specific environment to enable the AMRs to complete tasks in an efficient manner. This is a function th