Meego Wiki
Views

In-vehicle/Roadmap/FeatureScope

From MeeGo wiki
< In-vehicle | Roadmap(Difference between revisions)
Jump to: navigation, search
(Created page with " This topic provides an outline of requirements for a head unit. The requirements are derived from a real life development project for in-vehicle infotainment platform. ''This …")
(→UI_Speech_Recognition_Function)
 
(28 intermediate revisions not shown)
Line 1: Line 1:
-
 
This topic provides an outline of requirements for a head unit.  The requirements are derived from a real life development project for in-vehicle infotainment platform.  ''This is still work in progress--the content will be added gradually.''
This topic provides an outline of requirements for a head unit.  The requirements are derived from a real life development project for in-vehicle infotainment platform.  ''This is still work in progress--the content will be added gradually.''
Line 16: Line 15:
# After power-on, the head unit accepts first user input (e.g. HMI menu, screen, volume control) within max. 2 s.
# After power-on, the head unit accepts first user input (e.g. HMI menu, screen, volume control) within max. 2 s.
# After power-on, the head unit displays the Navigation map and starts the route guidance, resuming the previous route, within max. 5 s.
# After power-on, the head unit displays the Navigation map and starts the route guidance, resuming the previous route, within max. 5 s.
 +
# Head unit reacts to user input via HMI (touch screen events, button or switch presses) within max. 100 ms.  Head unit displays interim response (e.g. sand glass) for all requests that cannot be completed within this time
 +
# Head unit performs voice recognition for a single word against G2P vocabulary within max. 200 ms measured from the end of utterance (silence detection) to recognition result available (e.g. HMI display update or resulting voice prompt output).
 +
# Head unit performs hand writing recognition against Chinese standard 2000 character set within max. 200 ms measured from end of entry to recognition result available (e.g. HMI display update or resulting voice prompt output).
 +
# Head unit exhibits the refresh rate of min 15 fps for the navigation map in any scaling.
 +
# Head unit exhibits a graphics rendering frame rate (for HMI animations and Navigation maps) of min. 15 fps.
== Audio ==
== Audio ==
=== Audio_Codecs_Playback ===
=== Audio_Codecs_Playback ===
 +
# Head unit supports the following codecs:
 +
## AAC LC
 +
## MP3
 +
## WMA
 +
## WAV
 +
## Ogg Vorbis
 +
## FLAC
=== Audio_Processing_And_Management ===
=== Audio_Processing_And_Management ===
-
# up to 4 audio zones
+
# Head unit supports up to 4 independent audio zones (sinks).  One zone if 5.1 channels (e.g. cabin speakers) and others are stereo.  (Each zone might require its own independent audio processing and audio codec instances.)
-
# audio enhancements
+
# Head unit allows the user to adjust at least audio treble and bass level.
-
# any source to 4 sinks
+
# Head unit allows the user to adjust balance and fade level.
 +
# Head unit is capable of playing back any of its sources to any of 4 audio zones including simultaneous playback to up to 4 audio zones.
 +
# Head unit uses pre-defined fader ramps when switching the audio source for any particular audio zone.
 +
# For each audio zone, the head unit keeps track of the most recently played audio source and restores it upon startup.  (For example, when the FM tuner was played to the cabin speakers and the head unit has been switched off, it would attempt tuning to the same FM station and putting it to the cabin speakers upon the next startup.)
 +
# For each audio zone, all available audio sources are assigned a unique priority.  When an audio source is being played back and a higher priority source becomes available (e.g. the phone application attempts to play a ring tone), the latter overrides the former.
 +
## As a configuration option, overriding can be done either by muting the first source and playing back the second source, or by lowering the volume of the first source and mixing it with the second source.
=== Audio_Generation ===
=== Audio_Generation ===
Line 35: Line 51:
=== Audio_Volume_Control ===
=== Audio_Volume_Control ===
-
# vehicle speed dependent
+
# Volume level is adjustable by the user.
-
# volume level through shutdown cycle
+
## Adjustments by the user are only accepted during actual audio playback.  (In other words, any volume changes e.g. during start up phase and before the audio output can be heard will be ignored.)
 +
## The volume level selection by the user is maintained for each audio source / sink pair individually.
 +
# Volume level is adjusted by the head unit accordingly to the current vehicle speed.
 +
# Volume level set by the user is preserved through shutdown cycle.
 +
## Upon start up, the previous volume level is only restored up to a defined 'start up maximum' (as contrasted to the 'absolute maximum' that can be set after the head unit is up and running in its normal mode).
 +
# Under limit volume need to be set as some low level NOT minimum just to make sure that at least user can recognize low level volume even though user set the volume to minimum which is basically "MUTE" before the power cycle
 +
# Upper limit volume need to be set as some high level NOT maximum just to make sure that user can not be annoyed due to the previous maximum volume setting.
 +
# In Emergency Case, volume should not be adjusted by the user, volume need to be fit as some very loud level as much as no sound is masked with ambient noisy background,
 +
 
 +
=== Handfree Functionality ===
 +
# Noise reduction
 +
# Echo cancellation
 +
# Residual echo cancellation
 +
# Automatic gain control
 +
# Programmable Equalization
 +
# Performance Requirement
 +
  - objective testing : VDA 1.6
 +
  - subjective testing : In-Vehicle live testing
== Connectivity ==
== Connectivity ==
Line 84: Line 117:
==== UI_Touch_Sreen_Hand_Writing_Detection ====
==== UI_Touch_Sreen_Hand_Writing_Detection ====
==== UI_Speech_Recognition_Function ====
==== UI_Speech_Recognition_Function ====
 +
1.Recognition Category
 +
#discrete command - command, single digit ( e.g. "dial",'1','2',..etc )
 +
#continuous command - continuous digit ( e.g. "dial 555-1212", " 555-1212" )
 +
#natural language understaning - flexible recognition ( e.g. "I want to make a call to john","please route to the san franciso regency hyatt hotel",..etc)
 +
#Dynamic vocabulary (e.g. phonebook, MP3 Title)
 +
##number of phonebook and Music Title is not determined, User flexblly add and remove the list whatever they want, this list is dynamically loaded for a recognition, especially G2P (Grapheme-To-Phoneme) techinique is required to generate the phonetic transcription for the new words
 +
#VDE ( Voice Destination Entry )
 +
##Multi Step (e.g. "MI" and then "troy" and then "1307"
 +
##One shot (e.g. "1307 troy, MI")
 +
 +
2.Recognition Response time
 +
#discrete command - 300ms
 +
#continuous command - 1200ms
 +
#phone book & Music Title - 1200ms
 +
#natural language understanding, VDE - 1500ms
 +
 +
3.Recognition performance measurement
 +
#overall accuracy measurement
 +
##Average Sentence Accuracy = (Total Number of Correct Sentences)/(Total Number of Sentence Attempted)
 +
#individual accuracy measurement
 +
##Average Word Accuracy = ((Total number of attempts) - (insertions) - (deletions) - (substitutions))/Total number of attempts
 +
 
 +
4.Recognition performance requirement
 +
#discrete command
 +
##IDLE SNR>20dB (98%>) , Middle noisy SNR>10dB ( 95%> ), Too much noisy SNR> 6dB ( 92%) , SNR <6dB (rejection )
 +
#continuous command
 +
##IDLE SNR>20dB (98%>) , Middle noisy SNR>10dB ( 95%> ), Too much noisy SNR> 6dB ( 92%) , SNR <6dB (rejection )
 +
 +
5.Speech Recognition functionality
 +
#confidence score
 +
##the recognized result can be accepted/rejected according to confidence score, which represent how confidentially result can be accepted in terms of log likelihood, for example, if the developer set the confidence threshold to 40, basically assumed that confidence score range is between 0(low confidence score)to 100(high confidence score), result can be accepted only if confidence score is greater than confidence threshold 40
 +
#grammar weight
 +
##In case of poor accuracy command compare to other candidate in grammar, the weight can be adjusted to eqaulize the result, for example ( 1.1 dial | 0.9 store | 1.0 one | 1.0 two | 1.0 three | 0.9 four | .... | 1.1 oh ) , 1.1 means that more weight , and 1.0 is equal unity gain, 0.9 means that less weight compare to the unity gain
 +
#SNR rejection
 +
##if ambient noisy is too much, this can be measured by SNR (Signal to Noise Ratio), the it's better to reject recognized result in case of SNR is lower than some specific threahold which is potentially too much corrupted by noise condition
 +
##even though SNR is lower than some level which is not reliable to get the correct result. if confidence score is extremely high, the result can be accepted depend on the OEM specification
 +
#Talk too soon
 +
##if user start the utterance before the start beep is playback to user, there is tendancy to chop at the beginning of the utterance, so there is high possibility of result can be misrecognized, but this also can be accepted somewhat confidence score is too high to accepted the result
 +
#AGC(automatic Gain control)
 +
##speaking style is too difierent user by user, some user's voice is very low and smooth, and others are very strong and loud, even same user can speak different sound occasionally, so AGC is obviouly necessary to bring the volume up in case of soft voice, otherwise bringing the volume down against to loud voice. how this functionality also can be accepted/rejected according to their requirement. Usually AGC does not have impact to the accuracy. but recommed to use this functinality for higher usability.
 +
=== Multiple_HMI_Languages ===
=== Multiple_HMI_Languages ===
=== Software_Support ===
=== Software_Support ===
Line 99: Line 173:
== Navigation ==
== Navigation ==
=== Navigation_Engine ===
=== Navigation_Engine ===
 +
# map source
 +
# routing
 +
# searching (local / web service)
 +
# geocoding / reverse geocoding
 +
=== Map_Function ===
=== Map_Function ===
# database
# database
-
# 2D/3D representation
+
# 2D/3D representation (pitch)
 +
# zooming (IN / OUT / auto-zoom)
 +
# panning (NORTH / SOUTH / EAST / WEST / and combination of these 4 eg. NORTH-WEST)
 +
# orientation / northing (ON - map oriented north / OFF - map oriented in travelling direction)
 +
# follow gps signal ( ON - map cursor follows gps signal / OFF - map cursor does not follow gps signal / timeout - number of updates to wait before cursor follows gps signal on map; useful for panning)
 +
# set / clear destination
 +
# center the map
 +
# set/change map layout (day / night / detailed - POIs / plain simple)
 +
# bookmarks (see also Destination_Import)
 +
# OSD (on-screen display) information
 +
# set units: metric / imperial / ...
=== Real_Time_Traffic_Information ===
=== Real_Time_Traffic_Information ===
Line 108: Line 197:
=== Points_of_Interestes_POIs ===
=== Points_of_Interestes_POIs ===
=== Destination_Import ===
=== Destination_Import ===
 +
# from file (eg. USB)
 +
# from web server / web service
 +
# free input
 +
# coordinate units conversion
== Network ==
== Network ==
Line 173: Line 266:
# pre-installed
# pre-installed
# nomadic device
# nomadic device
 +
 +
[[Category:IVI]]

Latest revision as of 06:03, 28 June 2011

This topic provides an outline of requirements for a head unit. The requirements are derived from a real life development project for in-vehicle infotainment platform. This is still work in progress--the content will be added gradually.

Some of the requirements below will be fulfilled outside of the MeeGo IVI based software. For example, the implementation of CAN network interface and early audio functions most probably falls into this category. The decisions about implementation of specific requirements in MeeGo IVI software will be made assuming a specific system architecture.

Contents


Constraints

Performance

  1. After power-on, the head unit displays a static splash screen within 200 ms.
  2. After power-on, the head unit displays an animated splash screen (pre-defined, fixed) within 200 ms and at the frame rate above 4 fps.
  3. After power-on, the head unit is able to play early audio (e.g. chimes and gongs--minimal set, best suited type, simple) within max. 1 s.
  4. After power-on, the head unit plays FM radio, resuming the last station before shut-down, within max. 2 s.
  5. After power-on, the head unit plays compressed audio from a connected media device (e.g. MP3 from USB stick), resuming the last track played before shut-down, within max. 2 s.
  6. After power-on, the head unit displays the rear view camera output within max. 2 s.
  7. After power-on, the head unit accepts first user input (e.g. HMI menu, screen, volume control) within max. 2 s.
  8. After power-on, the head unit displays the Navigation map and starts the route guidance, resuming the previous route, within max. 5 s.
  9. Head unit reacts to user input via HMI (touch screen events, button or switch presses) within max. 100 ms. Head unit displays interim response (e.g. sand glass) for all requests that cannot be completed within this time
  10. Head unit performs voice recognition for a single word against G2P vocabulary within max. 200 ms measured from the end of utterance (silence detection) to recognition result available (e.g. HMI display update or resulting voice prompt output).
  11. Head unit performs hand writing recognition against Chinese standard 2000 character set within max. 200 ms measured from end of entry to recognition result available (e.g. HMI display update or resulting voice prompt output).
  12. Head unit exhibits the refresh rate of min 15 fps for the navigation map in any scaling.
  13. Head unit exhibits a graphics rendering frame rate (for HMI animations and Navigation maps) of min. 15 fps.

Audio

Audio_Codecs_Playback

  1. Head unit supports the following codecs:
    1. AAC LC
    2. MP3
    3. WMA
    4. WAV
    5. Ogg Vorbis
    6. FLAC

Audio_Processing_And_Management

  1. Head unit supports up to 4 independent audio zones (sinks). One zone if 5.1 channels (e.g. cabin speakers) and others are stereo. (Each zone might require its own independent audio processing and audio codec instances.)
  2. Head unit allows the user to adjust at least audio treble and bass level.
  3. Head unit allows the user to adjust balance and fade level.
  4. Head unit is capable of playing back any of its sources to any of 4 audio zones including simultaneous playback to up to 4 audio zones.
  5. Head unit uses pre-defined fader ramps when switching the audio source for any particular audio zone.
  6. For each audio zone, the head unit keeps track of the most recently played audio source and restores it upon startup. (For example, when the FM tuner was played to the cabin speakers and the head unit has been switched off, it would attempt tuning to the same FM station and putting it to the cabin speakers upon the next startup.)
  7. For each audio zone, all available audio sources are assigned a unique priority. When an audio source is being played back and a higher priority source becomes available (e.g. the phone application attempts to play a ring tone), the latter overrides the former.
    1. As a configuration option, overriding can be done either by muting the first source and playing back the second source, or by lowering the volume of the first source and mixing it with the second source.

Audio_Generation

  1. chimes and gongs
  2. calibration to interior

Audio_Recording

  1. voice
  2. radio
  3. podcasts

Audio_Volume_Control

  1. Volume level is adjustable by the user.
    1. Adjustments by the user are only accepted during actual audio playback. (In other words, any volume changes e.g. during start up phase and before the audio output can be heard will be ignored.)
    2. The volume level selection by the user is maintained for each audio source / sink pair individually.
  2. Volume level is adjusted by the head unit accordingly to the current vehicle speed.
  3. Volume level set by the user is preserved through shutdown cycle.
    1. Upon start up, the previous volume level is only restored up to a defined 'start up maximum' (as contrasted to the 'absolute maximum' that can be set after the head unit is up and running in its normal mode).
  4. Under limit volume need to be set as some low level NOT minimum just to make sure that at least user can recognize low level volume even though user set the volume to minimum which is basically "MUTE" before the power cycle
  5. Upper limit volume need to be set as some high level NOT maximum just to make sure that user can not be annoyed due to the previous maximum volume setting.
  6. In Emergency Case, volume should not be adjusted by the user, volume need to be fit as some very loud level as much as no sound is masked with ambient noisy background,

Handfree Functionality

  1. Noise reduction
  2. Echo cancellation
  3. Residual echo cancellation
  4. Automatic gain control
  5. Programmable Equalization
  6. Performance Requirement 
 - objective testing : VDA 1.6 
 - subjective testing : In-Vehicle live testing

Connectivity

Wireless_Interfaces

Wireless_Infrared_IRDA

Wireless_WiFi_Connection

  1. access point
  2. client

Wireless_Bluetooth_Connection_and_Telephony

Wireless_Cellular_Mobile_Telephony

Wireless_BT_Hands-free_Telephony

  1. profiles
  2. Acoustic Echo Cancellation / Noise Reduction (AEC / NR)
  3. single and dual microphone

Wired_Interfaces

USB_Connection_to_CE_Devices

Removable_SD_Card_Connections

Internet_Access_and_Browser

  1. internet radio
  2. podcasts
  3. applications (e.g. google maps)
  4. social networks

Phone_Functionality

  1. internal phone book
  2. VoIP

Terminal_Mode_Functionality

Microsoft_Plays_From_Device

Configuration

Personalization

  1. user profiles

Driver_Assistance

  1. camera input processing
  2. electronic horizon

Diagnostics

HMI

User_Inputs_Support

UI_Devices_Support

UI_Touch_Sreen_Hand_Writing_Detection

UI_Speech_Recognition_Function

1.Recognition Category

  1. discrete command - command, single digit ( e.g. "dial",'1','2',..etc )
  2. continuous command - continuous digit ( e.g. "dial 555-1212", " 555-1212" )
  3. natural language understaning - flexible recognition ( e.g. "I want to make a call to john","please route to the san franciso regency hyatt hotel",..etc)
  4. Dynamic vocabulary (e.g. phonebook, MP3 Title)
    1. number of phonebook and Music Title is not determined, User flexblly add and remove the list whatever they want, this list is dynamically loaded for a recognition, especially G2P (Grapheme-To-Phoneme) techinique is required to generate the phonetic transcription for the new words
  5. VDE ( Voice Destination Entry )
    1. Multi Step (e.g. "MI" and then "troy" and then "1307"
    2. One shot (e.g. "1307 troy, MI")

2.Recognition Response time

  1. discrete command - 300ms
  2. continuous command - 1200ms
  3. phone book & Music Title - 1200ms
  4. natural language understanding, VDE - 1500ms

3.Recognition performance measurement

  1. overall accuracy measurement
    1. Average Sentence Accuracy = (Total Number of Correct Sentences)/(Total Number of Sentence Attempted)
  2. individual accuracy measurement
    1. Average Word Accuracy = ((Total number of attempts) - (insertions) - (deletions) - (substitutions))/Total number of attempts

4.Recognition performance requirement

  1. discrete command
    1. IDLE SNR>20dB (98%>) , Middle noisy SNR>10dB ( 95%> ), Too much noisy SNR> 6dB ( 92%) , SNR <6dB (rejection )
  2. continuous command
    1. IDLE SNR>20dB (98%>) , Middle noisy SNR>10dB ( 95%> ), Too much noisy SNR> 6dB ( 92%) , SNR <6dB (rejection )

5.Speech Recognition functionality

  1. confidence score
    1. the recognized result can be accepted/rejected according to confidence score, which represent how confidentially result can be accepted in terms of log likelihood, for example, if the developer set the confidence threshold to 40, basically assumed that confidence score range is between 0(low confidence score)to 100(high confidence score), result can be accepted only if confidence score is greater than confidence threshold 40
  2. grammar weight
    1. In case of poor accuracy command compare to other candidate in grammar, the weight can be adjusted to eqaulize the result, for example ( 1.1 dial | 0.9 store | 1.0 one | 1.0 two | 1.0 three | 0.9 four | .... | 1.1 oh ) , 1.1 means that more weight , and 1.0 is equal unity gain, 0.9 means that less weight compare to the unity gain
  3. SNR rejection
    1. if ambient noisy is too much, this can be measured by SNR (Signal to Noise Ratio), the it's better to reject recognized result in case of SNR is lower than some specific threahold which is potentially too much corrupted by noise condition
    2. even though SNR is lower than some level which is not reliable to get the correct result. if confidence score is extremely high, the result can be accepted depend on the OEM specification
  4. Talk too soon
    1. if user start the utterance before the start beep is playback to user, there is tendancy to chop at the beginning of the utterance, so there is high possibility of result can be misrecognized, but this also can be accepted somewhat confidence score is too high to accepted the result
  5. AGC(automatic Gain control)
    1. speaking style is too difierent user by user, some user's voice is very low and smooth, and others are very strong and loud, even same user can speak different sound occasionally, so AGC is obviouly necessary to bring the volume up in case of soft voice, otherwise bringing the volume down against to loud voice. how this functionality also can be accepted/rejected according to their requirement. Usually AGC does not have impact to the accuracy. but recommed to use this functinality for higher usability.

Multiple_HMI_Languages

Software_Support

Users_Manual_Integrated

Graphics_Functions

Media

Media_Player_and_Browser_Requirements

  1. number of player instances
  2. playlists
  3. ripping

Storage_Media_Support

Navigation

Navigation_Engine

  1. map source
  2. routing
  3. searching (local / web service)
  4. geocoding / reverse geocoding

Map_Function

  1. database
  2. 2D/3D representation (pitch)
  3. zooming (IN / OUT / auto-zoom)
  4. panning (NORTH / SOUTH / EAST / WEST / and combination of these 4 eg. NORTH-WEST)
  5. orientation / northing (ON - map oriented north / OFF - map oriented in travelling direction)
  6. follow gps signal ( ON - map cursor follows gps signal / OFF - map cursor does not follow gps signal / timeout - number of updates to wait before cursor follows gps signal on map; useful for panning)
  7. set / clear destination
  8. center the map
  9. set/change map layout (day / night / detailed - POIs / plain simple)
  10. bookmarks (see also Destination_Import)
  11. OSD (on-screen display) information
  12. set units: metric / imperial / ...

Real_Time_Traffic_Information

Speech_Support_for_Navigation

Positioning_Capabilities

Points_of_Interestes_POIs

Destination_Import

  1. from file (eg. USB)
  2. from web server / web service
  3. free input
  4. coordinate units conversion

Network

CAN_Network_Interface

MOST_Network_Interface

Ethernet_Network_Interface

  1. DLNA
  2. AVB

Serial_Network_Interfaces

  1. UART
  2. LIN

Gateway

Power

Power_Management_Master

Power_Modes

Wakeup_Conditions

Monitor_Voltage_Ranges

Telematics and Data Services

Emergency_and_Breakdown_Call

JAVA_Telematics_Client

Remote_Diagnostics

Vehicle_Tracking

Radio and Video Broadcast Tuner

Radio_Broadcast_Reception

Radio_AM_FM_Broadcast_Reception

Radio_SDARS_Broadcast_Reception

Radio_HD_Broadcast_Reception

Radio_DAB_Broadcast_Reception

Radio_DRM_Broadcast_Reception

Broadcast_Sources_Pause_Resume

  1. time shift

Radio_Further_Traffic_and_Data_Services

Video_Broadcast_Reception

Video

Video_Playback_Formats_and_Codecs

  1. containers
  2. codecs
  3. resolutions
  4. frame rates

Video_Decoding_Performance

Voice

Voice_Tag_Processing

Voice_Recognition_Capabilties

Voice_Output_Functionality

Software Services

Software_Architectural_Requirements

Software_Update

Office_Applications

Security_Features

  1. user privacy
  2. data storage security
  3. data access authorization

Gaming_Features

  1. pre-installed
  2. nomadic device
Retrieved from "http://wiki.meego.com/In-vehicle/Roadmap/FeatureScope"
Personal tools