基于远程遥控方式训练松鼠识别毒品和爆炸物气味
Experiment of training squirrels to recognize the odor of drugs and explosives based on remote control
孙宁 1胡庭溪 2徐澍 1黄和 2刘慧颖 2尹劲 3俞曦 4潘彩霞 1赵建文 2徐杨 2许益维 2杨志东 1水映懿 5高志杰 1朱启文2
作者信息
- 1. 中国刑事警察学院,沈阳 110034
- 2. 辽宁省行为认知重点实验室,沈阳医学院,沈阳 110034
- 3. 重庆市合川区公安局,重庆 404100
- 4. 安徽省铜陵市公安局,安徽铜陵 244000
- 5. 大连外国语大学,大连 116000;中国人民公安大学,北京 100038
- 折叠
摘要
目的 松鼠作为一种新型生物探测器,尝试无人介入开展远程遥控训练其识别危险品气味.方法 本实验将远程可视智能喂食器与手机互联,登录手机APP后,用手机远程监测和遥控奖励方式对松鼠进行爆炸物与海洛因替代物气味的嗅认训练实验.结果 松鼠7天可迅速习得爆炸物气味,证实了松鼠的嗅觉敏锐能力和学习认知能力.在嗅认行为改进训练过程中,限于智能喂食器的技术局限,本实验没有成功建立松鼠嗅认气味的条件反射,而是建立了松鼠嗅认行为的条件反射.结论 实验结果表明行为的捕捉以及奖励的时机对松鼠行为监测与训练至关重要,提示采用自动化训练或遥控训练方法,需要先进的通信传输技术以及对松鼠生物学行为的细致捕捉和及时奖励.
Abstract
Objective As a new type of biological detector,squirrels are trained to recognize the smell of drugs and explosives by remote control without human intervention.Methods In this experiment,the remote visual smart feeder is connected to the mobile phone.Results After logging into the mobile phone APP,the squirrel is trained to smell the odor of explosives and heroin substitute using mobile phone remote monitoring and remote-control reward.Squirrels can quickly acquire the odor of explosives in 7 days,demonstrating their sensitive olfactory,intense learning,and cognition ability.Due to the technical limitations of the smart feeder,the squirrel's conditioned response to odor is not established by this method.However,the squirrel's conditional response to sniffing behavior is established randomly.Conclusion According to the results,the capture of behavior and the timing of rewards are very essential for squirrel behavior monitoring and training.Advanced communication and transmission technology,as well as meticulous capture of squirrel biological behavior and timely rewards are required for employing automatic training or remote-control training methods.
关键词
松鼠/训练/识别气味/爆炸物/海洛因替代物/远程遥控Key words
squirrel/training/recognizing the odor/explosives/heroin substitutes/mobile phone remote control引用本文复制引用
基金项目
国家重点研发计划"公共安全风险防控与应急技术装备"重点专项(2017YFC0803602)
出版年
2024
NETL
NSTL
万方数据