Component-based modeling of wireless mac protocols for efficient simulations
View Patent ↗Channel access delays and reception uncertainty are modeled as protocol-independent generic processes that are optimized for improved simulation performance. The generic process components are designed such that each different protocol can be modeled using an arrangement of these components that is specific to the protocol. In this way, speed and/or accuracy improvements to the generic process components are reflected in each of such protocol models. If an accurate analytic model is not available for the generic process component, a prediction engine, such as a neural network, is preferably used. The prediction engine is trained using the existing detailed models of network devices. Once trained, the prediction engine is used to model the generic process, and the protocol model that includes the generic component is used in lieu of the detailed models, thereby saving substantial processing time.
1. A network simulator comprising:
a set of components including:
a contention transmission component,
a reserved transmission component,
a contention reception component, and
a reserved reception component,
a plurality of device models including at least a first and second wireless MAC model,
a controller that is configured to control propagation of simulated traffic events in a network among at least some of the plurality of device models,
wherein:
the first wireless model includes the set of components arranged in a first protocol configuration, and
the second wireless model includes the set of components arranged in a second protocol configuration that differs from the first protocol configuration.
2. The network simulator of claim 1 , wherein one or more of the components of the set of components includes a prediction engine.
3. The network simulator of claim 2 , wherein the prediction engine includes a neural net.
4. The network simulator of claim 1 , wherein each of the contention and reserved transmission components is configured to provide an estimated channel access delay.
5. The network simulator of claim 1 , wherein the contention transmission component is configured to provide an estimated channel access delay, based upon:
a number of one-hop active nodes,
at least one packet size parameter,
at least one packet arrival parameter, and
at least one channel attempt parameter.
6. The network simulator of claim 1 , wherein the reserved transmission component is configured to provide an estimated channel access delay, based upon:
a number of one-hop active nodes,
at least one transmitting node queue parameter, and
at least one one-hop active node queue parameter.
7. The network simulator of claim 1 , wherein each of the contention and reserved reception components is configured to determine whether a transmitted packet is received successfully.
8. The network simulator of claim 1 , wherein the contention reception component is configured to determine whether a transmitted packet is received, based upon:
a number of one-hop active nodes,
at least one packet size parameter,
at least one packet arrival parameter, and
at least one signal distortion parameter.
9. The network simulator of claim 1 , wherein the reserved reception component is configured to determine whether a transmitted packet is received, based upon:
a number of non-participating non-one-hop active nodes, and
at least one signal distortion parameter.
10. The network simulator of claim 1 , wherein the controller is configured to train one or more components of the set of components based on a simulation of a third wireless MAC model that does not include the one or more components.
11. The network simulator of claim 10 , wherein the controller is configured to terminate the training of the one or more components and continue the simulation using the first MAC model.
12. The network simulator of claim 10 , wherein the controller is configured to train a plurality of variants of at least one component of the set of components, each variant corresponding to a particular region of the network.
13. A method of simulating a network, comprising:
propagating, on a network simulation machine, simulated traffic events among a plurality of device models that include at least a first and second wireless MAC model,
wherein
the propagating of the simulated traffic events at the first wireless MAC model includes executing a first sequence of transmissions and receptions,
the propagating of the simulated traffic events at the second wireless MAC model includes executing a second sequence of transmissions and receptions that differs significantly from the first sequence, and,
the method includes:
executing, on the network simulation machine, a common channel access process model to determine a first channel access delay for a transmission of the first sequence of transmissions,
executing, on the network simulation machine, the common channel access process model to determine a second channel access delay for a transmission of the second sequence of transmissions,
executing, on the network simulation machine, a common channel reception process model to determine a first likelihood of successful reception for a reception of the first sequence of receptions, and
executing, on the network simulation machine, the common channel reception process model to determine a second likelihood of successful reception for a reception of the second sequence of receptions.
14. The method of claim 13 , wherein the common channel access process model includes a contention transmission component and a reserved transmission component.
15. The method of claim 14 , wherein the common channel reception process model includes a contention reception component and a reserved reception component.
16. The method of claim 13 , wherein the common channel reception process model includes a contention reception component and a reserved reception component.
17. The method of claim 13 , wherein the common channel access process model is embodied as one or more neural nets.
18. The method of claim 13 , wherein the common channel reception process model is embodied as one or more neural nets.
19. The method of claim 13 , including training at least one model of the common channel access process model and the common access reception process model based on a simulation of a third wireless MAC model that does not include the at least one model.
20. The method of claim 19 , including:
terminating the training, and
continuing the simulation using the first MAC model.
21. The method of claim 19 , including training a plurality of variants of the at least one model.
22. A non-transient computer readable medium that includes a computer program that is configured to cause a processor to simulate a network, the computer program including program code that includes:
a set of components including:
a contention transmission component,
a reserved transmission component,
a contention reception component, and
a reserved reception component,
a plurality of device models including at least a first and second wireless MAC model,
wherein:
the first wireless model includes code corresponding to a first protocol that executes components of the set of components in a first predefined sequence, and
the second wireless model includes code corresponding to a second protocol that executes components of the set of components in a second predefined sequence that differs from the first predefined sequence.
23. The computer readable medium of claim 22 , wherein one or more of the components of the set of components is encoded as a neural net.
24. The computer readable medium of claim 22 , wherein the program code includes a training module that is configured to train at least one component of the set of components based on a simulation of the network using at least a third wireless MAC model that does not execute the at least one component.
25. The computer readable medium of claim 22 , wherein the training module is configured to train a plurality of variants of the at least one component, each variant corresponding to a particular region of the network.