明日とロボット

アンドロイドでArduinoロボット動かしてテレビとか操作してみた。

android

動機

New York Timesで、「2011年の終わりには、googleがアンドロイドフォンで家の電球を操作できるようになる。すごいね。」的なことを報道していたので、発表される前にやってみようかな。という軽い気持ちから。

目的

• アンドロイドフォンで家事用ロボット！
• 基本的なアイディアとしては、ロボットを遠隔操作して、そのロボットが見付けたデバイスを操作してみよう！みたいなもの。
• 折角なので、この間発表されたAndroid Open Accessory Development Kit (ADK)を使ってアンドロイドでArduino操作してみた。

概要

• デバイス（デモではテレビとLEDライト）にはそれぞれRFIDタグを割り振てられている
• ロボットにはRFIDリーダーが取り付けられていて、それぞれのデバイスのRFIDを読み取ることが出来る。
• 使うアンドロイドフォンは2台。
  • ロボットに載せてArduinoに命令を送りロボットコントロールするために一台（ロボットクライアント）
  • 遠隔地からそのロボットをコントロールするために一台（コントローラークライアント）
• ロボットはビデオイメージをコントローラにストリームバック。アンドロイドフォンをサーバーにして、外部から直接SDにアクセスさせる仕組み。
• ロボットが新しいデバイス（RFID情報）を取得する度にその情報と目的地のRFID情報をコントローラクライアントに送信
• コントローラクライアントは、その方法を受け取り次第データベースに参照し、新しいデバイスであれば操作するためのAPIプラグインをダウンロードインストール
• というわけで、どんなデバイスでも遠隔操作できるようになるという仕組み。

例

あ、部屋のアラーム留めるの忘れてた！
ルームメイトにあとでとやかく言われるのやだな。

しかたない。アンドロイドフォン使って留めとくか。

おっと、主から連絡が入ったぞ

よしよし、アラームクロックまで移動させられよう。

アラームクロックのRFIDをスキャンして。

RFID情報とMACアドレス情報をベースに、操作するためのインターフェースを取得。

よし、これで家のアラームを操作できる

電源、オフ！

といった感じ。。

デモ (with presentation)

＊デモは2分32秒から！

質問があればお気軽にー。

ツイートする

Permalink | コメント(0) | トラックバック(0) | 17:30

霊長類行動学16. グループ形成の理由

霊長類行動学

霊長類行動学もくじ

Scientific study

• create model
• develop hypothesis of the explanation
  • has to be falsifiable
• science cannot prove anything, but test the hypothesis
• hypothetical deductive framework

at the beginning of Primatology

• observational and deductive science
• status of Primatology was how much they had data
• people were interested in
  • 1. descriptive study:
    • details of behavior
  • 2. correlational study
    • ecological correlations of grouping patterns
• findings were based on correlational study

Crook (1972)

• provided intensive discussion
  • 1. "primate form groups because of predation risk"
  • 2. "sexual selection is very important"
• brought those discussion

Wrangham (1980)

• changed the way of study in Primatology
• introduced a hypothetical deductive framework
  • 1st principle is
    • 1. prediction
    • 2. test
• looked at relationship between behavior and ecology
• Wrangham did
  • 1. changes the way we classify social systems of primates
    • groups are not fixed entities
    • instead, they are a product of underlying patterns of male and female behavior
      • response to environment
  • 2. emphasized importance of understanding the explanatory model
    • now, primate became data to test hypothesis
  • 3. focuses on female as a driving force behind
    • group organization
    • thought to be that males solely play the key roles.
    • his arguments were too biased, but the idea opened up the question
    
    

defined the question

• "why they live in group?"
  • important question because it's odd among all other animals.
  • animals want to avoid competition so want to avoid grouping
  • for any animals, foraging by their own is better
    • group living should be disadvantages.
    • therefore, there should be some reasons
    • 1. living in group is actually an advantage
    • 2. cost stays, but getting other benefit more - less disadvantage

make the problem testable

• two different kinds of groups of females
  • 1. Kin
  • 2. non-kin
• Wrangham classified primates based on the type of group not based on group size
  • 1. "female bonded" - female stays in natal group
    • forms affiliative bonds with other females
  • 2. "not female bonded" - female leaves natal group
• establish the criteria to distinguish "female bonded" from "not female bonded"
  • 1. do female breed in natal group?
  • 2. do female transfer out of natal group?
  • 3. do female engage in affiliative actions
    • e.g., grooming, haddling, dominate, assistance, etc
• now, we can test hypothesis like if "Japanese macaque is female bonded"

why female live in group?

• cooperation must be an Evolutionary Stable Strategy (ESS)
  • he was creating a model, the mechanism
• rules:
  • suppose there is a fruit tree that can support only two individual
  • There exists a linear dominance hierarchy: A>B>C>D
• Condition 1: within dyadic competition (two fight with no support from others)
  • A, B on the tree, C, D on the ground
• Condition 2: any two individuals can cooperate, and they can defeat any other.
  • (how to form the cooperation tends to be based on kin.)
  • C+D on the tree, A, B on the ground, but soon
  • A+B on the tree, C+D on the ground
• Thus, grouping is needed for intergroup feeding competition.
• He argues that
• => grouping patterns based on environmental condition
  • if one exploit patchy, high quality resources
    • females bond
  • if one exploit disperse, low quality resources
    • no bonding
• => male attaches for resource defense
  • sexual selection is secondary for the reason to group
  • female behavior drives male dispersal
• but his discussion about males were mostly wrong.

	female-bonded	non female-bonded
importance of food patches in inter-group interactions	high	low
frequency of female participation in inter-group interactions	high	low
role of female in group more merit	active	passive
frequency of female-female grooming	high	low
frequency of male and male grooming in multi-male group	low	high

shift in the mainstream of the theory

• 1960's - 70's
  • female behavior is also important, but males are the most
• 1980's - females determine

after Wrangham

• Primatology is no longer a correlational study focusing on resource exploitation and resource availability
• emphasis on the predictable model.
• but where he failed was
  • he did not explain the primary question "why" primate form group, but only answerd "how".
  • idea about too much emphasis on sexual selection
  • he predicted that idea of predation drive grouping will be failed
• but inspired people to go out to test, and built the foundation of science in this field.

Covel van Schails

• falsifies food resource competition idea to explain the reason for grouping
• very strong hypothetical deductive framework
  • existence of alternative model
    • have to test the opponent idea as well
  • Wrangham's model had worked well only in folivore paradox

3 hypothesis for why primate live in groups

• 1. to avopid predation
  • increasing chance of defense
  • too costly for predator to attack
  • increase vigilance
• 2. predation defense + advantages of group living
  • help in locating and defending resources
  • regulate resource use
• 3. resource defense
  • predation is not an issue
  • regulation of group size
  
  

what makes group living bearable for individuals

• 1. food is in large patches - if a lot of food
• 2. individuals can only exploit limited amount of food
  • 1. ability to consume
  • 2. food itself is toxic (e.g., plant secondary compound), inadequate nutrition
• 3. different age and sex classes
  • use different food items
    • niche separation, niche dimorphism
    • e.g., orangutan, males and females eat different food
    
    

alternative reasons

• 1) animals regulate return time to patches
  • territory size increase, group size increase, defensibility decrease
  • => doesn't work
• 2) information exchange (like bees)
  • benefit from sharing information of food
  • food is dispersed in space and time
  • they will benefit
  
  

what lacks in his approach

• he made prediction and rejected all other ideas, but did not do for his model and predation model.

grouping v. population density

• 1. predation model' predict group size goes up
  • then, fall off as group split
• 2. intergroup feeding competition (IGFC)
  • group size should grow continuously
  • high population density -> more competition
  
  

groupsize v. fecundity (measured by #infants / female)

• 1. Predation:
  • fecundity falls within increasing group
    • higher competition -> low number of infants
• 2. IGFC
  • feeding increase with group size, then falls off as group get bigger
  
  

result

• data fit both predation models
• IGFC cannot explain why folivore is living in group

ツイートする

Permalink | コメント(0) | トラックバック(0) | 12:47