*5.2. Threat and Attack Models*

Existing attack models primarily make use of manual and single-point attacks targeting single measurement variables (sensor readings) or control commands. However, stealthy attacks, those trying to cause damage and at the same time remain undetected, may necessitate multi-point attacks if they are to evade detection mechanisms and operators. This area is starting to receive increased attention from researchers investigating the security of CPS [67]; however, more effort is required to understand how these attacks can be performed and what the limits on their effectiveness might be. Consequently, few studies have verified the effectiveness of existing detection models against these attacks.
