Skip to content

Latest commit

 

History

History
219 lines (178 loc) · 16.9 KB

README.md

File metadata and controls

219 lines (178 loc) · 16.9 KB
Raw

@awarns/human-activity

npm (scoped) npm

This module allows to detect and react to the changes in the activity being performed by the user (or object) carrying the phone: standing still, walking, running, riding a bicycle or being inside a vehicle. Depending on the granularity of the detection mechanism being used, it can even detect user stand-ups too.

This plugin acts as a wrapper on top of the nativescript-context-apis plugin (from the same authors), offering human activity change detection tasks. Obtain human activity updates even in background.

Install the plugin using the following command line instruction:

ns plugin add @awarns/human-activity

Usage

After installing and setting up this plugin, you'll have access to two task groups to start and stop listening human activity updates at different granularity levels. The received updates, will be a HumanActivityChange record, described below.

Setup

This plugin requires you to register its loader during the framework initialization, like this:

// ... platform imports
import { awarns } from '@awarns/core';
import { demoTasks } from '../tasks';
import { demoTaskGraph } from '../graph';
import { registerHumanActivityPlugin } from '@awarns/human-activity';

awarns
  .init(
    demoTasks,
    demoTaskGraph,
    [
      registerHumanActivityPlugin(),
    ]
  )
  // ... handle initialization promise

Plugin loader parameters:

This plugin loader takes no parameters.

Tasks

Task name Description
startDetectingCoarseHumanActivityChanges Allows to start detecting human activity changes at a coarse granularity level. This level of granularity means that activity changes will be reported no sooner than 1 minute after the activity started. In contrast, the detection mechanism will be more robust against in-between activity stops. This means, for example, that if the user is walking and stops for a few seconds, or is inside a vehicle and stops at traffic light, these subtle changes won't be detected at this granularity level
stopDetectingCoarseHumanActivityChanges The complement to the previous task. Allows to stop receiving coarse activity updates on demand
startDetectingIntermediateHumanActivityChanges Allows to start detecting human activity changes at an intermediate granularity level. This level of granularity means that activity changes will be reported as soon as they are detected. In contrast, the detection mechanism will be more sensitive to in-between activity stops. This means, for example, that if the user is walking and stops for a few seconds, or is inside a vehicle and stops at traffic light, these subtle changes will be detected as transitions to becoming still. Due to this extra of granularity, this activity detection mechanism is able to detect user stand-up actions (phone tilts abruptly)
stopDetectingIntermediateHumanActivityChanges The complement to the previous task. Allows to stop receiving intermediate activity updates on demand

Note: All the tasks require physical activity access permission for their execution. Each task will automatically request what is missing during framework's initialization

Start / stop receiving updates on human activity changes at a coarse level

To register these tasks for their use, you just need to import them and call their generator functions inside your application's task list:

import { Task } from '@awarns/core/tasks';
import {
  startDetectingCoarseHumanActivityChangesTask,
  stopDetectingCoarseHumanActivityChangesTask,
} from '@awarns/human-activity';

export const demoTasks: Array<Task> = [
  startDetectingCoarseHumanActivityChangesTask(),
  stopDetectingCoarseHumanActivityChangesTask(),
];

Task generator parameters:

These task generators take no parameters

Task output events:

These tasks don't produce significant events after they complete their execution aside from the regular {task-name}Finished events, which are: startDetectingCoarseHumanActivityChangesFinished and stopDetectingCoarseHumanActivityChangesFinished.

However, once the start task has finished running, relevant events will be emitted by the internal listeners once the corresponding action has been detected. These are listed below.

Example usage in the application task graph:

on('startEvent', run('startDetectingCoarseHumanActivityChanges'));
on('stopEvent', run('stopDetectingCoarseHumanActivityChanges'));

on('userActivityChanged', run('writeRecords'))

on('userStartedBeingStill', run('writeRecords'));
on('userFinishedBeingStill', run('writeRecords'));

on('userStartedWalking', run('writeRecords'));
on('userFinishedWalking', run('writeRecords'));

on('userStartedRunning', run('writeRecords'));
on('userFinishedRunning', run('writeRecords'));

on('userStartedRidingABicycle', run('writeRecords'));
on('userFinishedRidingABicycle', run('writeRecords'));

on('userStartedBeingInAVehicle', run('writeRecords'));
on('userFinishedBeingInAVehicle', run('writeRecords'));

Note: To use the writeRecords task, the persistence package must be installed and configured. See persistence package docs.

Start / stop receiving updates on human activity changes at an intermediate level

To register these tasks for their use, you just need to import them and call their generator functions inside your application's task list:

import { Task } from '@awarns/core/tasks';
import {
  startDetectingIntermediateHumanActivityChangesTask,
  stopDetectingIntermediateHumanActivityChangesTask,
} from '@awarns/human-activity';

export const demoTasks: Array<Task> = [
  startDetectingIntermediateHumanActivityChangesTask(),
  stopDetectingIntermediateHumanActivityChangesTask(),
];

Task generator parameters:

These task generators take no parameters

Task output events:

These tasks don't produce significant events after they complete their execution aside from the regular {task-name}Finished events, which are: startDetectingIntermediateHumanActivityChangesFinished and stopDetectingIntermediateHumanActivityChangesFinished.

However, once the start task has finished running, relevant events will be emitted by the internal listeners once the corresponding action has been detected. These are listed below.

Example usage in the application task graph:

on('startEvent', run('startDetectingCoarseHumanActivityChanges'));
on('stopEvent', run('stopDetectingCoarseHumanActivityChanges'));

on('userActivityChanged', run('writeRecords'))

on('userStartedBeingStill', run('writeRecords'));
on('userFinishedBeingStill', run('writeRecords'));

on('userStartedStandingUp', run('writeRecords'));
on('userFinishedStandingUp', run('writeRecords'));

on('userStartedWalking', run('writeRecords'));
on('userFinishedWalking', run('writeRecords'));

on('userStartedRunning', run('writeRecords'));
on('userFinishedRunning', run('writeRecords'));

on('userStartedRidingABicycle', run('writeRecords'));
on('userFinishedRidingABicycle', run('writeRecords'));

on('userStartedBeingInAVehicle', run('writeRecords'));
on('userFinishedBeingInAVehicle', run('writeRecords'));

Note: To use the writeRecords task, the persistence package must be installed and configured. See persistence package docs.

Events

Name Payload Description
userActivityChanged HumanActivityChange Indicates that the user activity has changed, either by ending an ongoing activity or starting a new one
userStartedBeingStill HumanActivityChange Indicates that the user of the phone has stopped moving
userFinishedBeingStill HumanActivityChange Indicates that the user of the phone has started moving
userStartedStandingUp HumanActivityChange Indicates that the user of the phone has started standing up
userFinishedStandingUp HumanActivityChange Indicates that the user of the phone has finished standing up
userStartedWalking HumanActivityChange Indicates that the user of the phone has started to walk
userFinishedWalking HumanActivityChange Indicates that the user of the phone has stopped walking
userStartedRunning HumanActivityChange Indicates that the user of the phone has started to run
userFinishedRunning HumanActivityChange Indicates that the user of the phone has stopped running
userStartedRidingABicycle HumanActivityChange Indicates that the user of the phone has started riding a bicycle
userFinishedRidingABicycle HumanActivityChange Indicates that the user of the phone has stopped riding a bicycle
userStartedBeingInAVehicle HumanActivityChange Indicates that the user of the phone has started being in a vehicle
userFinishedBeingInAVehicle HumanActivityChange Indicates that the user of the phone has stopped being in a vehicle

Records

HumanActivityChange

Property Type Description
id string Record's unique id
type string Always human-activity
change Change Can be either start or end. Indicates if the change reflects the activity starting or finishing
timestamp Date The local time when the change was detected
activity HumanActivity The activity which was detected. Can be any of the ones supported by the context-apis plugin (see context-apis API docs section on human activity, to obtain the full list)
confidence number Probability value of the detected action reported by the activity detection mechanism. Ranges from 0 to 1. This field will contain an undefined value in records reported by the coarse detection mechanism

License

Apache License Version 2.0